Model: gemini-3-pro-preview

sample_run06

Select Configuration:

📊 Sample Metrics

🔍 Exploration

last_node_coverage: 1
last_edge_coverage: 0.537
n_exploration_steps: 16
action_counts:
move: 5
rotate: 10
return: 0
observe: 15
term: 1
forced_term: 0
query: 0
action_cost: 15
final_information_gain: 0.897
action_fail_ratio: 0
valid_action_ratio: 1
false_belief_steps: 8
false_belief_f1: 1
false_belief_f1_position: 1
false_belief_f1_facing: 1
false_belief_action_cost: 7
false_belief_action_cost_after_seen: 0

✅ Evaluation

overall:
n_total: 27
total_score: 21.103
avg_accuracy: 0.782
per_task:
DirectionEvaluationTask:
n_total: 3
task_score: 2
avg_accuracy: 0.667
PovEvaluationTask:
n_total: 3
task_score: 2
avg_accuracy: 0.667
BackwardPovTextEvaluationTask:
n_total: 3
task_score: 2
avg_accuracy: 0.667
Action2ViewEvaluationTask:
n_total: 3
task_score: 3
avg_accuracy: 1
View2ActionTextEvaluationTask:
n_total: 3
task_score: 2
avg_accuracy: 0.667
AlloMappingEvaluationTask:
n_total: 3
task_score: 2.784
avg_accuracy: 0.928
RotEvaluationTask:
n_total: 3
task_score: 3
avg_accuracy: 1
Location2ViewEvaluationTask:
n_total: 3
task_score: 1.500
avg_accuracy: 0.500
View2LocationTextEvaluationTask:
n_total: 3
task_score: 2.819
avg_accuracy: 0.940

✅ Evaluation (prompt_cogmap)

overall:
n_total: 27
total_score: 21.177
avg_accuracy: 0.784
per_task:
DirectionEvaluationTask:
n_total: 3
task_score: 1.500
avg_accuracy: 0.500
PovEvaluationTask:
n_total: 3
task_score: 2.500
avg_accuracy: 0.833
BackwardPovTextEvaluationTask:
n_total: 3
task_score: 2
avg_accuracy: 0.667
Action2ViewEvaluationTask:
n_total: 3
task_score: 3
avg_accuracy: 1
View2ActionTextEvaluationTask:
n_total: 3
task_score: 1
avg_accuracy: 0.333
AlloMappingEvaluationTask:
n_total: 3
task_score: 2.677
avg_accuracy: 0.892
RotEvaluationTask:
n_total: 3
task_score: 3
avg_accuracy: 1
Location2ViewEvaluationTask:
n_total: 3
task_score: 2.500
avg_accuracy: 0.833
View2LocationTextEvaluationTask:
n_total: 3
task_score: 3
avg_accuracy: 1

✅ Evaluation (use_gt_cogmap)

overall:
n_total: 27
total_score: 27
avg_accuracy: 1
per_task:
DirectionEvaluationTask:
n_total: 3
task_score: 3
avg_accuracy: 1
PovEvaluationTask:
n_total: 3
task_score: 3
avg_accuracy: 1
BackwardPovTextEvaluationTask:
n_total: 3
task_score: 3
avg_accuracy: 1
Action2ViewEvaluationTask:
n_total: 3
task_score: 3
avg_accuracy: 1
View2ActionTextEvaluationTask:
n_total: 3
task_score: 3
avg_accuracy: 1
AlloMappingEvaluationTask:
n_total: 3
task_score: 3
avg_accuracy: 1
RotEvaluationTask:
n_total: 3
task_score: 3
avg_accuracy: 1
Location2ViewEvaluationTask:
n_total: 3
task_score: 3
avg_accuracy: 1
View2LocationTextEvaluationTask:
n_total: 3
task_score: 3
avg_accuracy: 1

✅ Evaluation (use_model_cogmap)

overall:
n_total: 27
total_score: 20.383
avg_accuracy: 0.755
per_task:
View2ActionTextEvaluationTask:
n_total: 3
task_score: 1
avg_accuracy: 0.333
AlloMappingEvaluationTask:
n_total: 3
task_score: 2.565
avg_accuracy: 0.855
RotEvaluationTask:
n_total: 3
task_score: 3
avg_accuracy: 1
Location2ViewEvaluationTask:
n_total: 3
task_score: 1.500
avg_accuracy: 0.500
View2LocationTextEvaluationTask:
n_total: 3
task_score: 2.819
avg_accuracy: 0.940
DirectionEvaluationTask:
n_total: 3
task_score: 1.500
avg_accuracy: 0.500
PovEvaluationTask:
n_total: 3
task_score: 2.500
avg_accuracy: 0.833
BackwardPovTextEvaluationTask:
n_total: 3
task_score: 3
avg_accuracy: 1
Action2ViewEvaluationTask:
n_total: 3
task_score: 2.500
avg_accuracy: 0.833

🧠 Cognitive Map

exploration:
error:
local_vs_gt_local_avg:
dir: 0.892
facing: 1
pos: 0.882
overall: 0.925
global_vs_gt_global_avg:
dir: 0.877
facing: 0.930
pos: 0.839
overall: 0.882
agent_vs_gt_agent_avg:
dir: 1
facing: 1
pos: 0.957
overall: 0.986
newly_observed_vs_gt_local_avg:
dir: 0.917
facing: 1
pos: 0.886
overall: 0.934
correctness:
last_global_vs_gt_full:
dir: 0.858
facing: 1
pos: 0.850
overall: 0.903
consistency:
local_vs_global_avg:
dir: 0.849
facing: 0.923
pos: 0.864
overall: 0.879
position_update_avg: 0.801
facing_update_avg: 0.953
position_stability_avg: 0.801
facing_stability_avg: 0.953
fog_probe:
f1_avg: 1
precision_avg: 1
recall_avg: 1
cogmap_fb:
metrics:
changed:
dir: None
pos: 0.668
facing: 1
overall: None
retention:
dir: None
pos: 0.256
facing: 0
overall: None
unchanged:
dir: 0.723
facing: 0.771
pos: 0.741
overall: 0.745
inertia: 0.016
inertia_list: [0.031128965472901334, 0.0011095130345293827]
unchanged_retention:
dir: None
pos: 0.803
facing: 0.625
overall: None
unchanged_retention_minus_retention:
pos: 0.898
facing: 1
unchanged_exploration:
dir: 0.873
facing: 1
pos: 0.844
overall: 0.906

🧭 False Belief CogMap

inertia: 0.016
changed:
dir: None
pos: 0.668
facing: 1
overall: None
retention:
dir: None
pos: 0.256
facing: 0
overall: None
unchanged:
dir: 0.723
facing: 0.771
pos: 0.741
overall: 0.745
unchanged_retention:
dir: None
pos: 0.803
facing: 0.625
overall: None
unchanged_retention_minus_retention:
pos: 0.898
facing: 1
unchanged_exploration:
dir: 0.873
facing: 1
pos: 0.844
overall: 0.906

📊 Performance Charts

Cognitive Map (Update)
Global Update Metrics
Cognitive Map (Full)
Global Full Metrics
Cognitive Map (Self-Tracking)
Global Self-Tracking Metrics
FB CogMap (Unchanged)
False Belief CogMap Unchanged per Turn
Fog Probe F1
Fog Probe F1 per Turn
Fog Probe Precision
Fog Probe Precision per Turn
Fog Probe Recall
Fog Probe Recall per Turn
Position Update
Position Update per Turn
Facing Update
Facing Update per Turn
Position Stability
Position Stability per Turn
Facing Stability
Facing Stability per Turn

🌍 Exploration Phase

🔄 Turn 1

💬 Assistant Action
Actions: [Observe()]
📊 Cognitive Map JSONs
🤖 Predicted
{
  "agent": {
    "position": [
      0,
      0
    ],
    "facing": "north"
  },
  "lego man": {
    "position": [
      -3,
      5
    ],
    "facing": "south"
  }
}
🎯 Ground Truth (Observed)
{
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  },
  "agent": {
    "position": [
      0,
      0
    ],
    "facing": "north"
  }
}
🎯 Ground Truth (Full)
{
  "bike": {
    "position": [
      -3,
      1
    ],
    "facing": "south"
  },
  "airplane": {
    "position": [
      3,
      -3
    ],
    "facing": "south"
  },
  "rubberduck": {
    "position": [
      -11,
      1
    ],
    "facing": "north"
  },
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  },
  "toilet": {
    "position": [
      7,
      -1
    ],
    "facing": "east"
  },
  "truck": {
    "position": [
      -7,
      1
    ],
    "facing": "east"
  },
  "backpack": {
    "position": [
      -1,
      -2
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      1,
      -2
    ],
    "facing": "west"
  },
  "fridge": {
    "position": [
      5,
      -3
    ],
    "facing": "east"
  },
  "soda can": {
    "position": [
      -7,
      3
    ],
    "facing": "east"
  },
  "basket": {
    "position": [
      5,
      1
    ],
    "facing": "east"
  },
  "lamp": {
    "position": [
      -11,
      5
    ],
    "facing": "east"
  },
  "red door": {
    "position": [
      -5,
      0
    ],
    "facing": "east"
  },
  "green door": {
    "position": [
      2,
      0
    ],
    "facing": "east"
  },
  "agent": {
    "position": [
      0,
      0
    ],
    "facing": "north"
  }
}
📊 Cognitive Map JSONs
🤖 Predicted
{
  "origin": "agent",
  "objects": {
    "lego man": {
      "position": [
        -3,
        5
      ],
      "facing": "-y"
    }
  }
}
🎯 Ground Truth
{
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  }
}
📊 Cognitive Map JSONs
🤖 Predicted
{
  "origin": "agent",
  "objects": {
    "lego man": {
      "position": [
        -3,
        5
      ],
      "facing": "-y"
    }
  }
}
🎯 Ground Truth
{
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  }
}
🗺️ Symbolic Fog Map
######################
#......###############
#......###############
#......#...B.E########
#......#A....D########
#......#......#......#
#......+....*.+......#
########.....C#......#
########F.....#......#
###############......#
###############......#
######################
🔍 Fog Probe JSONs
📍 Candidates: A, B, C, D, E, F
🤖 Predicted: A, C, F
🎯 Ground Truth: A, F, C
🧠 Cognitive Map Metrics
Global:
dir: 1
facing: 1
pos: 0.810
overall: 0.937
Global (Full):
dir: 0.025
facing: 0.222
pos: 0.108
overall: 0.118
Local:
dir: 1
facing: 1
pos: 0.743
overall: 0.914
Local (Newly):
dir: 1
facing: 1
pos: 0.743
overall: 0.914
Fog Probe:
precision: 1
recall: 1
overall: 1
📈 Turn Metrics
node_coverage: 0.071
edge_coverage: 0.024
is_action_fail: False
step: 1
action_counts:
move: 0
rotate: 0
return: 0
observe: 1
term: 0
forced_term: 0
query: 0
information_gain: 0.044
reward: -0.100
is_done: False
is_valid_action: True
Current state
State at Turn 1

🔄 Turn 2

💬 Assistant Action
Actions: [Rotate(90), Observe()]
📊 Cognitive Map JSONs
🤖 Predicted
{
  "agent": {
    "position": [
      0,
      0
    ],
    "facing": "east"
  },
  "lego man": {
    "position": [
      -3,
      5
    ],
    "facing": "south"
  },
  "green door": {
    "position": [
      2,
      0
    ]
  }
}
🎯 Ground Truth (Observed)
{
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  },
  "green door": {
    "position": [
      2,
      0
    ],
    "facing": "east"
  },
  "agent": {
    "position": [
      0,
      0
    ],
    "facing": "east"
  }
}
🎯 Ground Truth (Full)
{
  "bike": {
    "position": [
      -3,
      1
    ],
    "facing": "south"
  },
  "airplane": {
    "position": [
      3,
      -3
    ],
    "facing": "south"
  },
  "rubberduck": {
    "position": [
      -11,
      1
    ],
    "facing": "north"
  },
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  },
  "toilet": {
    "position": [
      7,
      -1
    ],
    "facing": "east"
  },
  "truck": {
    "position": [
      -7,
      1
    ],
    "facing": "east"
  },
  "backpack": {
    "position": [
      -1,
      -2
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      1,
      -2
    ],
    "facing": "west"
  },
  "fridge": {
    "position": [
      5,
      -3
    ],
    "facing": "east"
  },
  "soda can": {
    "position": [
      -7,
      3
    ],
    "facing": "east"
  },
  "basket": {
    "position": [
      5,
      1
    ],
    "facing": "east"
  },
  "lamp": {
    "position": [
      -11,
      5
    ],
    "facing": "east"
  },
  "red door": {
    "position": [
      -5,
      0
    ],
    "facing": "east"
  },
  "green door": {
    "position": [
      2,
      0
    ],
    "facing": "east"
  },
  "agent": {
    "position": [
      0,
      0
    ],
    "facing": "east"
  }
}
📊 Cognitive Map JSONs
🤖 Predicted
{
  "origin": "agent",
  "objects": {
    "green door": {
      "position": [
        0,
        2
      ]
    }
  }
}
🎯 Ground Truth
{
  "green door": {
    "position": [
      0,
      2
    ],
    "facing": "east"
  }
}
📊 Cognitive Map JSONs
🤖 Predicted
{
  "origin": "agent",
  "objects": {
    "green door": {
      "position": [
        0,
        2
      ]
    }
  }
}
🎯 Ground Truth
{
  "green door": {
    "position": [
      0,
      2
    ],
    "facing": "east"
  }
}
🗺️ Symbolic Fog Map
######################
#......###############
#......###############
#......#......########
#......#.A..F.########
#......#...E..#......#
#......+D...*C+......#
########....B.#......#
########......#......#
###############......#
###############......#
######################
🔍 Fog Probe JSONs
📍 Candidates: A, B, C, D, E, F
🤖 Predicted: A, B, D
🎯 Ground Truth: D, B, A
🧠 Cognitive Map Metrics
Global:
dir: 1
facing: 1
pos: 0.842
overall: 0.947
Global (Full):
dir: 0.050
facing: 0.222
pos: 0.168
overall: 0.147
Local:
dir: 1
facing: 1
pos: 1
overall: 1
Local (Newly):
dir: 1
facing: 1
pos: 1
overall: 1
Fog Probe:
precision: 1
recall: 1
overall: 1
📈 Turn Metrics
node_coverage: 0.143
edge_coverage: 0.049
is_action_fail: False
step: 2
action_counts:
move: 0
rotate: 1
return: 0
observe: 2
term: 0
forced_term: 0
query: 0
information_gain: 0.107
reward: -0.100
is_done: False
is_valid_action: True
Previous state
State before Turn 2
Current state
State at Turn 2

🔄 Turn 3

💬 Assistant Action
Actions: [Rotate(90), Observe()]
📊 Cognitive Map JSONs
🤖 Predicted
{
  "grid": {
    "width": 20,
    "height": 20
  },
  "agent": {
    "position": [
      0,
      0
    ],
    "facing": "south"
  },
  "lego man": {
    "position": [
      -3,
      5
    ],
    "facing": "south"
  },
  "green door": {
    "position": [
      2,
      0
    ],
    "facing": "east"
  },
  "backpack": {
    "position": [
      -2,
      -3
    ],
    "facing": "north"
  },
  "shoppingcart": {
    "position": [
      2,
      -3
    ],
    "facing": "east"
  }
}
🎯 Ground Truth (Observed)
{
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  },
  "backpack": {
    "position": [
      -1,
      -2
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      1,
      -2
    ],
    "facing": "west"
  },
  "green door": {
    "position": [
      2,
      0
    ],
    "facing": "east"
  },
  "agent": {
    "position": [
      0,
      0
    ],
    "facing": "south"
  }
}
🎯 Ground Truth (Full)
{
  "bike": {
    "position": [
      -3,
      1
    ],
    "facing": "south"
  },
  "airplane": {
    "position": [
      3,
      -3
    ],
    "facing": "south"
  },
  "rubberduck": {
    "position": [
      -11,
      1
    ],
    "facing": "north"
  },
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  },
  "toilet": {
    "position": [
      7,
      -1
    ],
    "facing": "east"
  },
  "truck": {
    "position": [
      -7,
      1
    ],
    "facing": "east"
  },
  "backpack": {
    "position": [
      -1,
      -2
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      1,
      -2
    ],
    "facing": "west"
  },
  "fridge": {
    "position": [
      5,
      -3
    ],
    "facing": "east"
  },
  "soda can": {
    "position": [
      -7,
      3
    ],
    "facing": "east"
  },
  "basket": {
    "position": [
      5,
      1
    ],
    "facing": "east"
  },
  "lamp": {
    "position": [
      -11,
      5
    ],
    "facing": "east"
  },
  "red door": {
    "position": [
      -5,
      0
    ],
    "facing": "east"
  },
  "green door": {
    "position": [
      2,
      0
    ],
    "facing": "east"
  },
  "agent": {
    "position": [
      0,
      0
    ],
    "facing": "south"
  }
}
📊 Cognitive Map JSONs
🤖 Predicted
{
  "origin": "agent",
  "objects": {
    "backpack": {
      "position": [
        2,
        3
      ],
      "facing": "+y"
    },
    "shoppingcart": {
      "position": [
        -2,
        3
      ],
      "facing": "+x"
    }
  }
}
🎯 Ground Truth
{
  "backpack": {
    "position": [
      1,
      2
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      -1,
      2
    ],
    "facing": "west"
  }
}
📊 Cognitive Map JSONs
🤖 Predicted
{
  "origin": "agent",
  "objects": {
    "backpack": {
      "position": [
        2,
        3
      ],
      "facing": "+y"
    },
    "shoppingcart": {
      "position": [
        -2,
        3
      ],
      "facing": "+x"
    }
  }
}
🎯 Ground Truth
{
  "backpack": {
    "position": [
      1,
      2
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      -1,
      2
    ],
    "facing": "west"
  }
}
🗺️ Symbolic Fog Map
######################
#......###############
#......###############
#......#.D....########
#......#.C.A.B########
#......#F.E...#......#
#......+....*.+......#
########......#......#
########......#......#
###############......#
###############......#
######################
🔍 Fog Probe JSONs
📍 Candidates: A, B, C, D, E, F
🤖 Predicted: C, E, F
🎯 Ground Truth: F, C, E
🧠 Cognitive Map Metrics
Global:
dir: 0.933
facing: 0.500
pos: 0.828
overall: 0.754
Global (Full):
dir: 0.117
facing: 0.222
pos: 0.276
overall: 0.205
Local:
dir: 1
facing: 1
pos: 0.810
overall: 0.937
Local (Newly):
dir: 1
facing: 1
pos: 0.810
overall: 0.937
Fog Probe:
precision: 1
recall: 1
overall: 1
📈 Turn Metrics
node_coverage: 0.286
edge_coverage: 0.098
is_action_fail: False
step: 3
action_counts:
move: 0
rotate: 2
return: 0
observe: 3
term: 0
forced_term: 0
query: 0
information_gain: 0.250
reward: -0.100
is_done: False
is_valid_action: True
Previous state
State before Turn 3
Current state
State at Turn 3

🔄 Turn 4

💬 Assistant Action
Actions: [Rotate(90), Observe()]
📊 Cognitive Map JSONs
🤖 Predicted
{
  "agent": {
    "position": [
      0,
      0
    ],
    "facing": "west"
  },
  "green door": {
    "position": [
      2,
      0
    ]
  },
  "red door": {
    "position": [
      -5,
      0
    ]
  },
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  },
  "bike": {
    "position": [
      -3,
      1
    ],
    "facing": "south"
  },
  "backpack": {
    "position": [
      -2,
      -2
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      2,
      -2
    ],
    "facing": "west"
  }
}
🎯 Ground Truth (Observed)
{
  "bike": {
    "position": [
      -3,
      1
    ],
    "facing": "south"
  },
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  },
  "backpack": {
    "position": [
      -1,
      -2
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      1,
      -2
    ],
    "facing": "west"
  },
  "red door": {
    "position": [
      -5,
      0
    ],
    "facing": "east"
  },
  "green door": {
    "position": [
      2,
      0
    ],
    "facing": "east"
  },
  "agent": {
    "position": [
      0,
      0
    ],
    "facing": "west"
  }
}
🎯 Ground Truth (Full)
{
  "bike": {
    "position": [
      -3,
      1
    ],
    "facing": "south"
  },
  "airplane": {
    "position": [
      3,
      -3
    ],
    "facing": "south"
  },
  "rubberduck": {
    "position": [
      -11,
      1
    ],
    "facing": "north"
  },
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  },
  "toilet": {
    "position": [
      7,
      -1
    ],
    "facing": "east"
  },
  "truck": {
    "position": [
      -7,
      1
    ],
    "facing": "east"
  },
  "backpack": {
    "position": [
      -1,
      -2
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      1,
      -2
    ],
    "facing": "west"
  },
  "fridge": {
    "position": [
      5,
      -3
    ],
    "facing": "east"
  },
  "soda can": {
    "position": [
      -7,
      3
    ],
    "facing": "east"
  },
  "basket": {
    "position": [
      5,
      1
    ],
    "facing": "east"
  },
  "lamp": {
    "position": [
      -11,
      5
    ],
    "facing": "east"
  },
  "red door": {
    "position": [
      -5,
      0
    ],
    "facing": "east"
  },
  "green door": {
    "position": [
      2,
      0
    ],
    "facing": "east"
  },
  "agent": {
    "position": [
      0,
      0
    ],
    "facing": "west"
  }
}
📊 Cognitive Map JSONs
🤖 Predicted
{
  "origin": "agent",
  "objects": {
    "bike": {
      "position": [
        1,
        3
      ],
      "facing": "-x"
    },
    "lego man": {
      "position": [
        3,
        5
      ],
      "facing": "-x"
    },
    "red door": {
      "position": [
        0,
        6
      ]
    }
  }
}
🎯 Ground Truth
{
  "red door": {
    "position": [
      0,
      5
    ],
    "facing": "east"
  },
  "bike": {
    "position": [
      1,
      3
    ],
    "facing": "south"
  },
  "lego man": {
    "position": [
      3,
      3
    ],
    "facing": "south"
  }
}
📊 Cognitive Map JSONs
🤖 Predicted
{
  "origin": "agent",
  "objects": {
    "bike": {
      "position": [
        1,
        3
      ],
      "facing": "-x"
    },
    "lego man": {
      "position": [
        3,
        5
      ],
      "facing": "-x"
    },
    "red door": {
      "position": [
        0,
        6
      ]
    }
  }
}
🎯 Ground Truth
{
  "red door": {
    "position": [
      0,
      5
    ],
    "facing": "east"
  },
  "bike": {
    "position": [
      1,
      3
    ],
    "facing": "south"
  }
}
🧠 Cognitive Map Metrics
Global:
dir: 0.929
facing: 1
pos: 0.924
overall: 0.951
Global (Full):
dir: 0.217
facing: 0.556
pos: 0.431
overall: 0.401
Local:
dir: 0.500
facing: 1
pos: 0.825
overall: 0.775
Local (Newly):
dir: 0.667
facing: 1
pos: 0.900
overall: 0.856
Fog Probe:
(none)
📈 Turn Metrics
node_coverage: 0.429
edge_coverage: 0.146
is_action_fail: False
step: 4
action_counts:
move: 0
rotate: 3
return: 0
observe: 4
term: 0
forced_term: 0
query: 0
information_gain: 0.405
reward: -0.100
is_done: False
is_valid_action: True
Previous state
State before Turn 4
Current state
State at Turn 4

🔄 Turn 5

💬 Assistant Action
Actions: [JumpTo(red door), Observe()]
📊 Cognitive Map JSONs
🤖 Predicted
{
  "agent": {
    "position": [
      -5,
      0
    ],
    "facing": "west"
  },
  "lego man": {
    "position": [
      -2,
      4
    ],
    "facing": "south"
  },
  "green door": {
    "position": [
      2,
      0
    ]
  },
  "backpack": {
    "position": [
      -2,
      -3
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      2,
      -3
    ],
    "facing": "west"
  },
  "bike": {
    "position": [
      -3,
      1
    ],
    "facing": "south"
  },
  "red door": {
    "position": [
      -5,
      0
    ]
  },
  "rubberduck": {
    "position": [
      -9,
      1
    ],
    "facing": "north"
  },
  "truck": {
    "position": [
      -7,
      2
    ],
    "facing": "east"
  },
  "lamp": {
    "position": [
      -8,
      4
    ]
  }
}
🎯 Ground Truth (Observed)
{
  "bike": {
    "position": [
      -3,
      1
    ],
    "facing": "south"
  },
  "rubberduck": {
    "position": [
      -11,
      1
    ],
    "facing": "north"
  },
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  },
  "truck": {
    "position": [
      -7,
      1
    ],
    "facing": "east"
  },
  "backpack": {
    "position": [
      -1,
      -2
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      1,
      -2
    ],
    "facing": "west"
  },
  "lamp": {
    "position": [
      -11,
      5
    ],
    "facing": "east"
  },
  "red door": {
    "position": [
      -5,
      0
    ],
    "facing": "east"
  },
  "green door": {
    "position": [
      2,
      0
    ],
    "facing": "east"
  },
  "agent": {
    "position": [
      -5,
      0
    ],
    "facing": "west"
  }
}
🎯 Ground Truth (Full)
{
  "bike": {
    "position": [
      -3,
      1
    ],
    "facing": "south"
  },
  "airplane": {
    "position": [
      3,
      -3
    ],
    "facing": "south"
  },
  "rubberduck": {
    "position": [
      -11,
      1
    ],
    "facing": "north"
  },
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  },
  "toilet": {
    "position": [
      7,
      -1
    ],
    "facing": "east"
  },
  "truck": {
    "position": [
      -7,
      1
    ],
    "facing": "east"
  },
  "backpack": {
    "position": [
      -1,
      -2
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      1,
      -2
    ],
    "facing": "west"
  },
  "fridge": {
    "position": [
      5,
      -3
    ],
    "facing": "east"
  },
  "soda can": {
    "position": [
      -7,
      3
    ],
    "facing": "east"
  },
  "basket": {
    "position": [
      5,
      1
    ],
    "facing": "east"
  },
  "lamp": {
    "position": [
      -11,
      5
    ],
    "facing": "east"
  },
  "red door": {
    "position": [
      -5,
      0
    ],
    "facing": "east"
  },
  "green door": {
    "position": [
      2,
      0
    ],
    "facing": "east"
  },
  "agent": {
    "position": [
      -5,
      0
    ],
    "facing": "west"
  }
}
📊 Cognitive Map JSONs
🤖 Predicted
{
  "origin": "agent",
  "objects": {
    "rubberduck": {
      "position": [
        2,
        6
      ],
      "facing": "+x"
    },
    "truck": {
      "position": [
        2,
        3
      ],
      "facing": "-y"
    },
    "lamp": {
      "position": [
        4,
        5
      ]
    }
  }
}
🎯 Ground Truth
{
  "truck": {
    "position": [
      1,
      2
    ],
    "facing": "east"
  },
  "lamp": {
    "position": [
      5,
      6
    ],
    "facing": "east"
  },
  "rubberduck": {
    "position": [
      1,
      6
    ],
    "facing": "north"
  }
}
📊 Cognitive Map JSONs
🤖 Predicted
{
  "origin": "agent",
  "objects": {
    "rubberduck": {
      "position": [
        2,
        6
      ],
      "facing": "+x"
    },
    "truck": {
      "position": [
        2,
        3
      ],
      "facing": "-y"
    },
    "lamp": {
      "position": [
        4,
        5
      ]
    }
  }
}
🎯 Ground Truth
{
  "truck": {
    "position": [
      1,
      2
    ],
    "facing": "east"
  },
  "lamp": {
    "position": [
      5,
      6
    ],
    "facing": "east"
  },
  "rubberduck": {
    "position": [
      1,
      6
    ],
    "facing": "north"
  }
}
🗺️ Symbolic Fog Map
######################
#..E...###############
#......###############
#....CF#......########
#......#......########
#D..B..#......#......#
#..A...*......+......#
########......#......#
########......#......#
###############......#
###############......#
######################
🔍 Fog Probe JSONs
📍 Candidates: A, B, C, D, E, F
🤖 Predicted: C, E, F
🎯 Ground Truth: C, F, E
🧠 Cognitive Map Metrics
Global:
dir: 0.836
facing: 1
pos: 0.806
overall: 0.881
Global (Full):
dir: 0.383
facing: 0.778
pos: 0.537
overall: 0.566
Local:
dir: 0.833
facing: 1
pos: 0.825
overall: 0.886
Local (Newly):
dir: 0.833
facing: 1
pos: 0.825
overall: 0.886
Fog Probe:
precision: 1
recall: 1
overall: 1
📈 Turn Metrics
node_coverage: 0.643
edge_coverage: 0.220
is_action_fail: False
step: 5
action_counts:
move: 1
rotate: 3
return: 0
observe: 5
term: 0
forced_term: 0
query: 0
information_gain: 0.528
reward: -0.100
is_done: False
is_valid_action: True
Previous state
State before Turn 5
Current state
State at Turn 5

🔄 Turn 6

💬 Assistant Action
Actions: [JumpTo(truck), Observe()]
📊 Cognitive Map JSONs
🤖 Predicted
{
  "agent": {
    "position": [
      -7,
      2
    ],
    "facing": "west"
  },
  "lego man": {
    "position": [
      -3,
      4
    ],
    "facing": "south"
  },
  "green door": {
    "position": [
      2,
      0
    ]
  },
  "backpack": {
    "position": [
      -2,
      -2
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      2,
      -2
    ],
    "facing": "west"
  },
  "bike": {
    "position": [
      -3,
      1
    ],
    "facing": "south"
  },
  "red door": {
    "position": [
      -5,
      0
    ]
  },
  "truck": {
    "position": [
      -7,
      2
    ],
    "facing": "east"
  },
  "rubberduck": {
    "position": [
      -10,
      2
    ],
    "facing": "north"
  },
  "lamp": {
    "position": [
      -9,
      5
    ]
  }
}
🎯 Ground Truth (Observed)
{
  "bike": {
    "position": [
      -3,
      1
    ],
    "facing": "south"
  },
  "rubberduck": {
    "position": [
      -11,
      1
    ],
    "facing": "north"
  },
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  },
  "truck": {
    "position": [
      -7,
      1
    ],
    "facing": "east"
  },
  "backpack": {
    "position": [
      -1,
      -2
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      1,
      -2
    ],
    "facing": "west"
  },
  "lamp": {
    "position": [
      -11,
      5
    ],
    "facing": "east"
  },
  "red door": {
    "position": [
      -5,
      0
    ],
    "facing": "east"
  },
  "green door": {
    "position": [
      2,
      0
    ],
    "facing": "east"
  },
  "agent": {
    "position": [
      -7,
      1
    ],
    "facing": "west"
  }
}
🎯 Ground Truth (Full)
{
  "bike": {
    "position": [
      -3,
      1
    ],
    "facing": "south"
  },
  "airplane": {
    "position": [
      3,
      -3
    ],
    "facing": "south"
  },
  "rubberduck": {
    "position": [
      -11,
      1
    ],
    "facing": "north"
  },
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  },
  "toilet": {
    "position": [
      7,
      -1
    ],
    "facing": "east"
  },
  "truck": {
    "position": [
      -7,
      1
    ],
    "facing": "east"
  },
  "backpack": {
    "position": [
      -1,
      -2
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      1,
      -2
    ],
    "facing": "west"
  },
  "fridge": {
    "position": [
      5,
      -3
    ],
    "facing": "east"
  },
  "soda can": {
    "position": [
      -7,
      3
    ],
    "facing": "east"
  },
  "basket": {
    "position": [
      5,
      1
    ],
    "facing": "east"
  },
  "lamp": {
    "position": [
      -11,
      5
    ],
    "facing": "east"
  },
  "red door": {
    "position": [
      -5,
      0
    ],
    "facing": "east"
  },
  "green door": {
    "position": [
      2,
      0
    ],
    "facing": "east"
  },
  "agent": {
    "position": [
      -7,
      1
    ],
    "facing": "west"
  }
}
📊 Cognitive Map JSONs
🤖 Predicted
{
  "origin": "agent",
  "objects": {
    "rubberduck": {
      "position": [
        0,
        3
      ],
      "facing": "+x"
    },
    "lamp": {
      "position": [
        4,
        4
      ]
    }
  }
}
🎯 Ground Truth
{
  "lamp": {
    "position": [
      4,
      4
    ],
    "facing": "east"
  },
  "rubberduck": {
    "position": [
      0,
      4
    ],
    "facing": "north"
  }
}
📊 Cognitive Map JSONs
🤖 Predicted
{
  "origin": "agent",
  "objects": {
    "rubberduck": {
      "position": [
        0,
        3
      ],
      "facing": "+x"
    },
    "lamp": {
      "position": [
        4,
        4
      ]
    }
  }
}
🎯 Ground Truth
(no data)
🗺️ Symbolic Fog Map
######################
#....A.###############
#..B...###############
#..E..D#......########
#.....F#......########
#..C.*.#......#......#
#......+......+......#
########......#......#
########......#......#
###############......#
###############......#
######################
🔍 Fog Probe JSONs
📍 Candidates: A, B, C, D, E, F
🤖 Predicted: A, D, F
🎯 Ground Truth: F, D, A
🧠 Cognitive Map Metrics
Global:
dir: 0.891
facing: 1
pos: 0.856
overall: 0.915
Global (Full):
dir: 0.408
facing: 0.778
pos: 0.570
overall: 0.585
Local:
dir: 1
facing: 1
pos: 0.900
overall: 0.967
Local (Newly):
(none)
Fog Probe:
precision: 1
recall: 1
overall: 1
📈 Turn Metrics
node_coverage: 0.643
edge_coverage: 0.268
is_action_fail: False
step: 6
action_counts:
move: 2
rotate: 3
return: 0
observe: 6
term: 0
forced_term: 0
query: 0
information_gain: 0.535
reward: -0.100
is_done: False
is_valid_action: True
Previous state
State before Turn 6
Current state
State at Turn 6

🔄 Turn 7

💬 Assistant Action
Actions: [Rotate(90), Observe()]
📊 Cognitive Map JSONs
🤖 Predicted
{
  "agent": {
    "position": [
      -7,
      2
    ],
    "facing": "north"
  },
  "lego man": {
    "position": [
      -3,
      4
    ],
    "facing": "south"
  },
  "green door": {
    "position": [
      2,
      0
    ]
  },
  "backpack": {
    "position": [
      -2,
      -3
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      2,
      -3
    ],
    "facing": "west"
  },
  "bike": {
    "position": [
      -4,
      1
    ],
    "facing": "south"
  },
  "red door": {
    "position": [
      -5,
      0
    ]
  },
  "truck": {
    "position": [
      -7,
      2
    ],
    "facing": "east"
  },
  "rubberduck": {
    "position": [
      -10,
      2
    ],
    "facing": "north"
  },
  "lamp": {
    "position": [
      -9,
      4
    ]
  },
  "soda can": {
    "position": [
      -7,
      4
    ]
  }
}
🎯 Ground Truth (Observed)
{
  "bike": {
    "position": [
      -3,
      1
    ],
    "facing": "south"
  },
  "rubberduck": {
    "position": [
      -11,
      1
    ],
    "facing": "north"
  },
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  },
  "truck": {
    "position": [
      -7,
      1
    ],
    "facing": "east"
  },
  "backpack": {
    "position": [
      -1,
      -2
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      1,
      -2
    ],
    "facing": "west"
  },
  "soda can": {
    "position": [
      -7,
      3
    ],
    "facing": "east"
  },
  "lamp": {
    "position": [
      -11,
      5
    ],
    "facing": "east"
  },
  "red door": {
    "position": [
      -5,
      0
    ],
    "facing": "east"
  },
  "green door": {
    "position": [
      2,
      0
    ],
    "facing": "east"
  },
  "agent": {
    "position": [
      -7,
      1
    ],
    "facing": "north"
  }
}
🎯 Ground Truth (Full)
{
  "bike": {
    "position": [
      -3,
      1
    ],
    "facing": "south"
  },
  "airplane": {
    "position": [
      3,
      -3
    ],
    "facing": "south"
  },
  "rubberduck": {
    "position": [
      -11,
      1
    ],
    "facing": "north"
  },
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  },
  "toilet": {
    "position": [
      7,
      -1
    ],
    "facing": "east"
  },
  "truck": {
    "position": [
      -7,
      1
    ],
    "facing": "east"
  },
  "backpack": {
    "position": [
      -1,
      -2
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      1,
      -2
    ],
    "facing": "west"
  },
  "fridge": {
    "position": [
      5,
      -3
    ],
    "facing": "east"
  },
  "soda can": {
    "position": [
      -7,
      3
    ],
    "facing": "east"
  },
  "basket": {
    "position": [
      5,
      1
    ],
    "facing": "east"
  },
  "lamp": {
    "position": [
      -11,
      5
    ],
    "facing": "east"
  },
  "red door": {
    "position": [
      -5,
      0
    ],
    "facing": "east"
  },
  "green door": {
    "position": [
      2,
      0
    ],
    "facing": "east"
  },
  "agent": {
    "position": [
      -7,
      1
    ],
    "facing": "north"
  }
}
📊 Cognitive Map JSONs
🤖 Predicted
{
  "origin": "agent",
  "objects": {
    "soda can": {
      "position": [
        0,
        2
      ]
    },
    "lamp": {
      "position": [
        -4,
        4
      ]
    }
  }
}
🎯 Ground Truth
{
  "lamp": {
    "position": [
      -4,
      4
    ],
    "facing": "east"
  },
  "soda can": {
    "position": [
      0,
      2
    ],
    "facing": "east"
  }
}
📊 Cognitive Map JSONs
🤖 Predicted
{
  "origin": "agent",
  "objects": {
    "soda can": {
      "position": [
        0,
        2
      ]
    },
    "lamp": {
      "position": [
        -4,
        4
      ]
    }
  }
}
🎯 Ground Truth
{
  "soda can": {
    "position": [
      0,
      2
    ],
    "facing": "east"
  }
}
🧠 Cognitive Map Metrics
Global:
dir: 0.864
facing: 1
pos: 0.836
overall: 0.900
Global (Full):
dir: 0.475
facing: 0.778
pos: 0.613
overall: 0.622
Local:
dir: 1
facing: 1
pos: 1
overall: 1
Local (Newly):
dir: 1
facing: 1
pos: 1
overall: 1
Fog Probe:
(none)
📈 Turn Metrics
node_coverage: 0.714
edge_coverage: 0.293
is_action_fail: False
step: 7
action_counts:
move: 2
rotate: 4
return: 0
observe: 7
term: 0
forced_term: 0
query: 0
information_gain: 0.581
reward: -0.100
is_done: False
is_valid_action: True
Previous state
State before Turn 7
Current state
State at Turn 7

🔄 Turn 8

💬 Assistant Action
Actions: [Rotate(90), Observe()]
📊 Cognitive Map JSONs
🤖 Predicted
{
  "agent": {
    "position": [
      -7,
      2
    ],
    "facing": "east"
  },
  "green door": {
    "position": [
      2,
      0
    ]
  },
  "red door": {
    "position": [
      -5,
      0
    ]
  },
  "lego man": {
    "position": [
      -3,
      4
    ],
    "facing": "south"
  },
  "bike": {
    "position": [
      -3,
      1
    ],
    "facing": "south"
  },
  "backpack": {
    "position": [
      -2,
      -3
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      2,
      -3
    ],
    "facing": "west"
  },
  "truck": {
    "position": [
      -7,
      2
    ],
    "facing": "east"
  },
  "soda can": {
    "position": [
      -7,
      4
    ]
  },
  "rubberduck": {
    "position": [
      -10,
      2
    ],
    "facing": "north"
  },
  "lamp": {
    "position": [
      -11,
      5
    ]
  }
}
🎯 Ground Truth (Observed)
{
  "bike": {
    "position": [
      -3,
      1
    ],
    "facing": "south"
  },
  "rubberduck": {
    "position": [
      -11,
      1
    ],
    "facing": "north"
  },
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  },
  "truck": {
    "position": [
      -7,
      1
    ],
    "facing": "east"
  },
  "backpack": {
    "position": [
      -1,
      -2
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      1,
      -2
    ],
    "facing": "west"
  },
  "soda can": {
    "position": [
      -7,
      3
    ],
    "facing": "east"
  },
  "lamp": {
    "position": [
      -11,
      5
    ],
    "facing": "east"
  },
  "red door": {
    "position": [
      -5,
      0
    ],
    "facing": "east"
  },
  "green door": {
    "position": [
      2,
      0
    ],
    "facing": "east"
  },
  "agent": {
    "position": [
      -7,
      1
    ],
    "facing": "east"
  }
}
🎯 Ground Truth (Full)
{
  "bike": {
    "position": [
      -3,
      1
    ],
    "facing": "south"
  },
  "airplane": {
    "position": [
      3,
      -3
    ],
    "facing": "south"
  },
  "rubberduck": {
    "position": [
      -11,
      1
    ],
    "facing": "north"
  },
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  },
  "toilet": {
    "position": [
      7,
      -1
    ],
    "facing": "east"
  },
  "truck": {
    "position": [
      -7,
      1
    ],
    "facing": "east"
  },
  "backpack": {
    "position": [
      -1,
      -2
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      1,
      -2
    ],
    "facing": "west"
  },
  "fridge": {
    "position": [
      5,
      -3
    ],
    "facing": "east"
  },
  "soda can": {
    "position": [
      -7,
      3
    ],
    "facing": "east"
  },
  "basket": {
    "position": [
      5,
      1
    ],
    "facing": "east"
  },
  "lamp": {
    "position": [
      -11,
      5
    ],
    "facing": "east"
  },
  "red door": {
    "position": [
      -5,
      0
    ],
    "facing": "east"
  },
  "green door": {
    "position": [
      2,
      0
    ],
    "facing": "east"
  },
  "agent": {
    "position": [
      -7,
      1
    ],
    "facing": "east"
  }
}
📊 Cognitive Map JSONs
🤖 Predicted
{
  "origin": "agent",
  "objects": {
    "red door": {
      "position": [
        2,
        3
      ],
      "facing": "-y"
    }
  }
}
🎯 Ground Truth
{
  "red door": {
    "position": [
      1,
      2
    ],
    "facing": "east"
  }
}
📊 Cognitive Map JSONs
🤖 Predicted
{
  "origin": "agent",
  "objects": {
    "red door": {
      "position": [
        2,
        3
      ],
      "facing": "-y"
    }
  }
}
🎯 Ground Truth
(no data)
🧠 Cognitive Map Metrics
Global:
dir: 0.864
facing: 1
pos: 0.868
overall: 0.910
Global (Full):
dir: 0.475
facing: 0.778
pos: 0.636
overall: 0.630
Local:
dir: 1
facing: 1
pos: 0.810
overall: 0.937
Local (Newly):
(none)
Fog Probe:
(none)
📈 Turn Metrics
node_coverage: 0.714
edge_coverage: 0.293
is_action_fail: False
step: 8
action_counts:
move: 2
rotate: 5
return: 0
observe: 8
term: 0
forced_term: 0
query: 0
information_gain: 0.581
reward: -0.100
is_done: False
is_valid_action: True
Previous state
State before Turn 8
Current state
State at Turn 8

🔄 Turn 9

💬 Assistant Action
Actions: [Rotate(90), Observe()]
📊 Cognitive Map JSONs
🤖 Predicted
{
  "agent": {
    "position": [
      -7,
      2
    ],
    "facing": "south"
  },
  "green door": {
    "position": [
      2,
      0
    ],
    "facing": "east"
  },
  "red door": {
    "position": [
      -5,
      0
    ],
    "facing": "west"
  },
  "lego man": {
    "position": [
      -4,
      4
    ],
    "facing": "south"
  },
  "bike": {
    "position": [
      -6,
      4
    ],
    "facing": "south"
  },
  "backpack": {
    "position": [
      -2,
      -3
    ],
    "facing": "north"
  },
  "shoppingcart": {
    "position": [
      2,
      -3
    ],
    "facing": "east"
  },
  "truck": {
    "position": [
      -7,
      2
    ],
    "facing": "east"
  },
  "rubberduck": {
    "position": [
      -10,
      2
    ],
    "facing": "north"
  },
  "lamp": {
    "position": [
      -9,
      5
    ]
  },
  "soda can": {
    "position": [
      -7,
      3
    ]
  }
}
🎯 Ground Truth (Observed)
{
  "bike": {
    "position": [
      -3,
      1
    ],
    "facing": "south"
  },
  "rubberduck": {
    "position": [
      -11,
      1
    ],
    "facing": "north"
  },
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  },
  "truck": {
    "position": [
      -7,
      1
    ],
    "facing": "east"
  },
  "backpack": {
    "position": [
      -1,
      -2
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      1,
      -2
    ],
    "facing": "west"
  },
  "soda can": {
    "position": [
      -7,
      3
    ],
    "facing": "east"
  },
  "lamp": {
    "position": [
      -11,
      5
    ],
    "facing": "east"
  },
  "red door": {
    "position": [
      -5,
      0
    ],
    "facing": "east"
  },
  "green door": {
    "position": [
      2,
      0
    ],
    "facing": "east"
  },
  "agent": {
    "position": [
      -7,
      1
    ],
    "facing": "south"
  }
}
🎯 Ground Truth (Full)
{
  "bike": {
    "position": [
      -3,
      1
    ],
    "facing": "south"
  },
  "airplane": {
    "position": [
      3,
      -3
    ],
    "facing": "south"
  },
  "rubberduck": {
    "position": [
      -11,
      1
    ],
    "facing": "north"
  },
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  },
  "toilet": {
    "position": [
      7,
      -1
    ],
    "facing": "east"
  },
  "truck": {
    "position": [
      -7,
      1
    ],
    "facing": "east"
  },
  "backpack": {
    "position": [
      -1,
      -2
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      1,
      -2
    ],
    "facing": "west"
  },
  "fridge": {
    "position": [
      5,
      -3
    ],
    "facing": "east"
  },
  "soda can": {
    "position": [
      -7,
      3
    ],
    "facing": "east"
  },
  "basket": {
    "position": [
      5,
      1
    ],
    "facing": "east"
  },
  "lamp": {
    "position": [
      -11,
      5
    ],
    "facing": "east"
  },
  "red door": {
    "position": [
      -5,
      0
    ],
    "facing": "east"
  },
  "green door": {
    "position": [
      2,
      0
    ],
    "facing": "east"
  },
  "agent": {
    "position": [
      -7,
      1
    ],
    "facing": "south"
  }
}
📊 Cognitive Map JSONs
🤖 Predicted
{
  "origin": "agent",
  "objects": {}
}
🎯 Ground Truth
(no data)
📊 Cognitive Map JSONs
🤖 Predicted
{
  "origin": "agent",
  "objects": {}
}
🎯 Ground Truth
(no data)
🧠 Cognitive Map Metrics
Global:
dir: 0.727
facing: 0.714
pos: 0.776
overall: 0.739
Global (Full):
dir: 0.400
facing: 0.556
pos: 0.569
overall: 0.508
Local:
(none)
Local (Newly):
(none)
Fog Probe:
(none)
📈 Turn Metrics
node_coverage: 0.714
edge_coverage: 0.293
is_action_fail: False
step: 9
action_counts:
move: 2
rotate: 6
return: 0
observe: 9
term: 0
forced_term: 0
query: 0
information_gain: 0.581
reward: -0.100
is_done: False
is_valid_action: True
Previous state
State before Turn 9
Current state
State at Turn 9

🔄 Turn 10

💬 Assistant Action
Actions: [Rotate(-90), JumpTo(red door), Observe()]
📊 Cognitive Map JSONs
🤖 Predicted
{
  "agent": {
    "position": [
      -5,
      0
    ],
    "facing": "east"
  },
  "lego man": {
    "position": [
      -3,
      5
    ],
    "facing": "south"
  },
  "green door": {
    "position": [
      2,
      0
    ]
  },
  "backpack": {
    "position": [
      -2,
      -3
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      2,
      -3
    ],
    "facing": "west"
  },
  "bike": {
    "position": [
      -3,
      1
    ],
    "facing": "south"
  },
  "red door": {
    "position": [
      -5,
      0
    ]
  },
  "truck": {
    "position": [
      -7,
      2
    ],
    "facing": "east"
  },
  "rubberduck": {
    "position": [
      -9,
      2
    ],
    "facing": "north"
  },
  "lamp": {
    "position": [
      -9,
      6
    ]
  },
  "soda can": {
    "position": [
      -7,
      4
    ]
  }
}
🎯 Ground Truth (Observed)
{
  "bike": {
    "position": [
      -3,
      1
    ],
    "facing": "south"
  },
  "rubberduck": {
    "position": [
      -11,
      1
    ],
    "facing": "north"
  },
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  },
  "truck": {
    "position": [
      -7,
      1
    ],
    "facing": "east"
  },
  "backpack": {
    "position": [
      -1,
      -2
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      1,
      -2
    ],
    "facing": "west"
  },
  "soda can": {
    "position": [
      -7,
      3
    ],
    "facing": "east"
  },
  "lamp": {
    "position": [
      -11,
      5
    ],
    "facing": "east"
  },
  "red door": {
    "position": [
      -5,
      0
    ],
    "facing": "east"
  },
  "green door": {
    "position": [
      2,
      0
    ],
    "facing": "east"
  },
  "agent": {
    "position": [
      -5,
      0
    ],
    "facing": "east"
  }
}
🎯 Ground Truth (Full)
{
  "bike": {
    "position": [
      -3,
      1
    ],
    "facing": "south"
  },
  "airplane": {
    "position": [
      3,
      -3
    ],
    "facing": "south"
  },
  "rubberduck": {
    "position": [
      -11,
      1
    ],
    "facing": "north"
  },
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  },
  "toilet": {
    "position": [
      7,
      -1
    ],
    "facing": "east"
  },
  "truck": {
    "position": [
      -7,
      1
    ],
    "facing": "east"
  },
  "backpack": {
    "position": [
      -1,
      -2
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      1,
      -2
    ],
    "facing": "west"
  },
  "fridge": {
    "position": [
      5,
      -3
    ],
    "facing": "east"
  },
  "soda can": {
    "position": [
      -7,
      3
    ],
    "facing": "east"
  },
  "basket": {
    "position": [
      5,
      1
    ],
    "facing": "east"
  },
  "lamp": {
    "position": [
      -11,
      5
    ],
    "facing": "east"
  },
  "red door": {
    "position": [
      -5,
      0
    ],
    "facing": "east"
  },
  "green door": {
    "position": [
      2,
      0
    ],
    "facing": "east"
  },
  "agent": {
    "position": [
      -5,
      0
    ],
    "facing": "east"
  }
}
📊 Cognitive Map JSONs
🤖 Predicted
{
  "origin": "agent",
  "objects": {
    "green door": {
      "position": [
        0,
        6
      ],
      "facing": "-y"
    },
    "bike": {
      "position": [
        -2,
        2
      ],
      "facing": "+x"
    },
    "backpack": {
      "position": [
        4,
        4
      ],
      "facing": "+x"
    },
    "shoppingcart": {
      "position": [
        1,
        6
      ],
      "facing": "-y"
    }
  }
}
🎯 Ground Truth
{
  "backpack": {
    "position": [
      2,
      4
    ],
    "facing": "south"
  },
  "green door": {
    "position": [
      0,
      7
    ],
    "facing": "east"
  },
  "bike": {
    "position": [
      -1,
      2
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      2,
      6
    ],
    "facing": "west"
  }
}
📊 Cognitive Map JSONs
🤖 Predicted
{
  "origin": "agent",
  "objects": {
    "green door": {
      "position": [
        0,
        6
      ],
      "facing": "-y"
    },
    "bike": {
      "position": [
        -2,
        2
      ],
      "facing": "+x"
    },
    "backpack": {
      "position": [
        4,
        4
      ],
      "facing": "+x"
    },
    "shoppingcart": {
      "position": [
        1,
        6
      ],
      "facing": "-y"
    }
  }
}
🎯 Ground Truth
(no data)
🧠 Cognitive Map Metrics
Global:
dir: 0.788
facing: 1
pos: 0.818
overall: 0.869
Global (Full):
dir: 0.433
facing: 0.778
pos: 0.600
overall: 0.604
Local:
dir: 0.600
facing: 1
pos: 0.821
overall: 0.807
Local (Newly):
(none)
Fog Probe:
(none)
📈 Turn Metrics
node_coverage: 0.714
edge_coverage: 0.390
is_action_fail: False
step: 10
action_counts:
move: 3
rotate: 7
return: 0
observe: 10
term: 0
forced_term: 0
query: 0
information_gain: 0.644
reward: -0.100
is_done: False
is_valid_action: True
Previous state
State before Turn 10
Current state
State at Turn 10

🔄 Turn 11

💬 Assistant Action
Actions: [JumpTo(green door), Observe()]
📊 Cognitive Map JSONs
🤖 Predicted
{
  "agent": {
    "position": [
      2,
      0
    ],
    "facing": "east"
  },
  "lego man": {
    "position": [
      -2,
      2
    ],
    "facing": "south"
  },
  "green door": {
    "position": [
      2,
      0
    ]
  },
  "backpack": {
    "position": [
      -2,
      -2
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      2,
      -2
    ],
    "facing": "west"
  },
  "bike": {
    "position": [
      -4,
      2
    ],
    "facing": "south"
  },
  "red door": {
    "position": [
      -5,
      0
    ]
  },
  "truck": {
    "position": [
      -7,
      2
    ],
    "facing": "east"
  },
  "rubberduck": {
    "position": [
      -10,
      2
    ],
    "facing": "north"
  },
  "lamp": {
    "position": [
      -10,
      4
    ]
  },
  "soda can": {
    "position": [
      -7,
      3
    ]
  },
  "toilet": {
    "position": [
      7,
      -1
    ],
    "facing": "east"
  },
  "fridge": {
    "position": [
      5,
      -3
    ]
  },
  "basket": {
    "position": [
      5,
      1
    ]
  }
}
🎯 Ground Truth (Observed)
{
  "bike": {
    "position": [
      -3,
      1
    ],
    "facing": "south"
  },
  "rubberduck": {
    "position": [
      -11,
      1
    ],
    "facing": "north"
  },
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  },
  "toilet": {
    "position": [
      7,
      -1
    ],
    "facing": "east"
  },
  "truck": {
    "position": [
      -7,
      1
    ],
    "facing": "east"
  },
  "backpack": {
    "position": [
      -1,
      -2
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      1,
      -2
    ],
    "facing": "west"
  },
  "fridge": {
    "position": [
      5,
      -3
    ],
    "facing": "east"
  },
  "soda can": {
    "position": [
      -7,
      3
    ],
    "facing": "east"
  },
  "basket": {
    "position": [
      5,
      1
    ],
    "facing": "east"
  },
  "lamp": {
    "position": [
      -11,
      5
    ],
    "facing": "east"
  },
  "red door": {
    "position": [
      -5,
      0
    ],
    "facing": "east"
  },
  "green door": {
    "position": [
      2,
      0
    ],
    "facing": "east"
  },
  "agent": {
    "position": [
      2,
      0
    ],
    "facing": "east"
  }
}
🎯 Ground Truth (Full)
{
  "bike": {
    "position": [
      -3,
      1
    ],
    "facing": "south"
  },
  "airplane": {
    "position": [
      3,
      -3
    ],
    "facing": "south"
  },
  "rubberduck": {
    "position": [
      -11,
      1
    ],
    "facing": "north"
  },
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  },
  "toilet": {
    "position": [
      7,
      -1
    ],
    "facing": "east"
  },
  "truck": {
    "position": [
      -7,
      1
    ],
    "facing": "east"
  },
  "backpack": {
    "position": [
      -1,
      -2
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      1,
      -2
    ],
    "facing": "west"
  },
  "fridge": {
    "position": [
      5,
      -3
    ],
    "facing": "east"
  },
  "soda can": {
    "position": [
      -7,
      3
    ],
    "facing": "east"
  },
  "basket": {
    "position": [
      5,
      1
    ],
    "facing": "east"
  },
  "lamp": {
    "position": [
      -11,
      5
    ],
    "facing": "east"
  },
  "red door": {
    "position": [
      -5,
      0
    ],
    "facing": "east"
  },
  "green door": {
    "position": [
      2,
      0
    ],
    "facing": "east"
  },
  "agent": {
    "position": [
      2,
      0
    ],
    "facing": "east"
  }
}
📊 Cognitive Map JSONs
🤖 Predicted
{
  "origin": "agent",
  "objects": {
    "toilet": {
      "position": [
        1,
        6
      ],
      "facing": "+y"
    },
    "fridge": {
      "position": [
        4,
        5
      ]
    },
    "basket": {
      "position": [
        -1,
        3
      ]
    }
  }
}
🎯 Ground Truth
{
  "toilet": {
    "position": [
      1,
      5
    ],
    "facing": "east"
  },
  "fridge": {
    "position": [
      3,
      3
    ],
    "facing": "east"
  },
  "basket": {
    "position": [
      -1,
      3
    ],
    "facing": "east"
  }
}
📊 Cognitive Map JSONs
🤖 Predicted
{
  "origin": "agent",
  "objects": {
    "toilet": {
      "position": [
        1,
        6
      ],
      "facing": "+y"
    },
    "fridge": {
      "position": [
        4,
        5
      ]
    },
    "basket": {
      "position": [
        -1,
        3
      ]
    }
  }
}
🎯 Ground Truth
{
  "toilet": {
    "position": [
      1,
      5
    ],
    "facing": "east"
  },
  "fridge": {
    "position": [
      3,
      3
    ],
    "facing": "east"
  },
  "basket": {
    "position": [
      -1,
      3
    ],
    "facing": "east"
  }
}
🗺️ Symbolic Fog Map
######################
#......###############
#......###############
#......#......########
#......#......########
#......#......#......#
#......+......*......#
########......#C..F..#
########......#B...A.#
###############.D....#
###############E.....#
######################
🔍 Fog Probe JSONs
📍 Candidates: A, B, C, D, E, F
🤖 Predicted: B, D, E
🎯 Ground Truth: D, E, B
🧠 Cognitive Map Metrics
Global:
dir: 0.876
facing: 1
pos: 0.876
overall: 0.918
Global (Full):
dir: 0.767
facing: 0.889
pos: 0.818
overall: 0.825
Local:
dir: 0.833
facing: 1
pos: 0.810
overall: 0.881
Local (Newly):
dir: 0.833
facing: 1
pos: 0.810
overall: 0.881
Fog Probe:
precision: 1
recall: 1
overall: 1
📈 Turn Metrics
node_coverage: 0.929
edge_coverage: 0.463
is_action_fail: False
step: 11
action_counts:
move: 4
rotate: 7
return: 0
observe: 11
term: 0
forced_term: 0
query: 0
information_gain: 0.793
reward: -0.100
is_done: False
is_valid_action: True
Previous state
State before Turn 11
Current state
State at Turn 11

🔄 Turn 12

💬 Assistant Action
Actions: [JumpTo(basket), Observe()]
📊 Cognitive Map JSONs
🤖 Predicted
{
  "agent": {
    "position": [
      5,
      1
    ],
    "facing": "east"
  },
  "bike": {
    "position": [
      -4,
      2
    ],
    "facing": "south"
  },
  "lego man": {
    "position": [
      -2,
      2
    ],
    "facing": "south"
  },
  "red door": {
    "position": [
      -4,
      0
    ],
    "facing": "east"
  },
  "green door": {
    "position": [
      2,
      0
    ],
    "facing": "west"
  },
  "backpack": {
    "position": [
      -2,
      -3
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      2,
      -3
    ],
    "facing": "west"
  },
  "truck": {
    "position": [
      -6,
      2
    ],
    "facing": "east"
  },
  "rubberduck": {
    "position": [
      -8,
      2
    ],
    "facing": "north"
  },
  "lamp": {
    "position": [
      -8,
      4
    ]
  },
  "soda can": {
    "position": [
      -6,
      4
    ]
  },
  "basket": {
    "position": [
      5,
      1
    ]
  },
  "toilet": {
    "position": [
      6,
      -1
    ],
    "facing": "east"
  },
  "fridge": {
    "position": [
      5,
      -3
    ]
  }
}
🎯 Ground Truth (Observed)
{
  "bike": {
    "position": [
      -3,
      1
    ],
    "facing": "south"
  },
  "rubberduck": {
    "position": [
      -11,
      1
    ],
    "facing": "north"
  },
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  },
  "toilet": {
    "position": [
      7,
      -1
    ],
    "facing": "east"
  },
  "truck": {
    "position": [
      -7,
      1
    ],
    "facing": "east"
  },
  "backpack": {
    "position": [
      -1,
      -2
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      1,
      -2
    ],
    "facing": "west"
  },
  "fridge": {
    "position": [
      5,
      -3
    ],
    "facing": "east"
  },
  "soda can": {
    "position": [
      -7,
      3
    ],
    "facing": "east"
  },
  "basket": {
    "position": [
      5,
      1
    ],
    "facing": "east"
  },
  "lamp": {
    "position": [
      -11,
      5
    ],
    "facing": "east"
  },
  "red door": {
    "position": [
      -5,
      0
    ],
    "facing": "east"
  },
  "green door": {
    "position": [
      2,
      0
    ],
    "facing": "east"
  },
  "agent": {
    "position": [
      5,
      1
    ],
    "facing": "east"
  }
}
🎯 Ground Truth (Full)
{
  "bike": {
    "position": [
      -3,
      1
    ],
    "facing": "south"
  },
  "airplane": {
    "position": [
      3,
      -3
    ],
    "facing": "south"
  },
  "rubberduck": {
    "position": [
      -11,
      1
    ],
    "facing": "north"
  },
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  },
  "toilet": {
    "position": [
      7,
      -1
    ],
    "facing": "east"
  },
  "truck": {
    "position": [
      -7,
      1
    ],
    "facing": "east"
  },
  "backpack": {
    "position": [
      -1,
      -2
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      1,
      -2
    ],
    "facing": "west"
  },
  "fridge": {
    "position": [
      5,
      -3
    ],
    "facing": "east"
  },
  "soda can": {
    "position": [
      -7,
      3
    ],
    "facing": "east"
  },
  "basket": {
    "position": [
      5,
      1
    ],
    "facing": "east"
  },
  "lamp": {
    "position": [
      -11,
      5
    ],
    "facing": "east"
  },
  "red door": {
    "position": [
      -5,
      0
    ],
    "facing": "east"
  },
  "green door": {
    "position": [
      2,
      0
    ],
    "facing": "east"
  },
  "agent": {
    "position": [
      5,
      1
    ],
    "facing": "east"
  }
}
📊 Cognitive Map JSONs
🤖 Predicted
{
  "origin": "agent",
  "objects": {
    "toilet": {
      "position": [
        2,
        2
      ],
      "facing": "+y"
    }
  }
}
🎯 Ground Truth
{
  "toilet": {
    "position": [
      2,
      2
    ],
    "facing": "east"
  }
}
📊 Cognitive Map JSONs
🤖 Predicted
{
  "origin": "agent",
  "objects": {
    "toilet": {
      "position": [
        2,
        2
      ],
      "facing": "+y"
    }
  }
}
🎯 Ground Truth
(no data)
🗺️ Symbolic Fog Map
######################
#......###############
#......###############
#......#......########
#......#......########
#......#......#.B*...#
#......+......+..F...#
########......#......#
########......#......#
###############ED..C.#
###############..A...#
######################
🔍 Fog Probe JSONs
📍 Candidates: A, B, C, D, E, F
🤖 Predicted: A, D, E
🎯 Ground Truth: A, D, E
🧠 Cognitive Map Metrics
Global:
dir: 0.838
facing: 1
pos: 0.793
overall: 0.877
Global (Full):
dir: 0.733
facing: 0.889
pos: 0.740
overall: 0.787
Local:
dir: 1
facing: 1
pos: 1
overall: 1
Local (Newly):
(none)
Fog Probe:
precision: 1
recall: 1
overall: 1
📈 Turn Metrics
node_coverage: 0.929
edge_coverage: 0.488
is_action_fail: False
step: 12
action_counts:
move: 5
rotate: 7
return: 0
observe: 12
term: 0
forced_term: 0
query: 0
information_gain: 0.806
reward: -0.100
is_done: False
is_valid_action: True
Previous state
State before Turn 12
Current state
State at Turn 12

🔄 Turn 13

💬 Assistant Action
Actions: [Rotate(90), Observe()]
📊 Cognitive Map JSONs
🤖 Predicted
{
  "agent": {
    "position": [
      5,
      1
    ],
    "facing": "south"
  },
  "green door": {
    "position": [
      2,
      0
    ]
  },
  "red door": {
    "position": [
      -5,
      0
    ]
  },
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  },
  "bike": {
    "position": [
      -3,
      2
    ],
    "facing": "south"
  },
  "backpack": {
    "position": [
      -1,
      -2
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      1,
      -2
    ],
    "facing": "west"
  },
  "truck": {
    "position": [
      -7,
      2
    ],
    "facing": "east"
  },
  "rubberduck": {
    "position": [
      -9,
      2
    ],
    "facing": "north"
  },
  "soda can": {
    "position": [
      -7,
      3
    ]
  },
  "lamp": {
    "position": [
      -9,
      4
    ]
  },
  "basket": {
    "position": [
      5,
      1
    ]
  },
  "fridge": {
    "position": [
      5,
      -2
    ]
  },
  "toilet": {
    "position": [
      7,
      -1
    ],
    "facing": "east"
  },
  "airplane": {
    "position": [
      3,
      -2
    ],
    "facing": "south"
  }
}
🎯 Ground Truth (Observed)
{
  "bike": {
    "position": [
      -3,
      1
    ],
    "facing": "south"
  },
  "airplane": {
    "position": [
      3,
      -3
    ],
    "facing": "south"
  },
  "rubberduck": {
    "position": [
      -11,
      1
    ],
    "facing": "north"
  },
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  },
  "toilet": {
    "position": [
      7,
      -1
    ],
    "facing": "east"
  },
  "truck": {
    "position": [
      -7,
      1
    ],
    "facing": "east"
  },
  "backpack": {
    "position": [
      -1,
      -2
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      1,
      -2
    ],
    "facing": "west"
  },
  "fridge": {
    "position": [
      5,
      -3
    ],
    "facing": "east"
  },
  "soda can": {
    "position": [
      -7,
      3
    ],
    "facing": "east"
  },
  "basket": {
    "position": [
      5,
      1
    ],
    "facing": "east"
  },
  "lamp": {
    "position": [
      -11,
      5
    ],
    "facing": "east"
  },
  "red door": {
    "position": [
      -5,
      0
    ],
    "facing": "east"
  },
  "green door": {
    "position": [
      2,
      0
    ],
    "facing": "east"
  },
  "agent": {
    "position": [
      5,
      1
    ],
    "facing": "south"
  }
}
🎯 Ground Truth (Full)
{
  "bike": {
    "position": [
      -3,
      1
    ],
    "facing": "south"
  },
  "airplane": {
    "position": [
      3,
      -3
    ],
    "facing": "south"
  },
  "rubberduck": {
    "position": [
      -11,
      1
    ],
    "facing": "north"
  },
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  },
  "toilet": {
    "position": [
      7,
      -1
    ],
    "facing": "east"
  },
  "truck": {
    "position": [
      -7,
      1
    ],
    "facing": "east"
  },
  "backpack": {
    "position": [
      -1,
      -2
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      1,
      -2
    ],
    "facing": "west"
  },
  "fridge": {
    "position": [
      5,
      -3
    ],
    "facing": "east"
  },
  "soda can": {
    "position": [
      -7,
      3
    ],
    "facing": "east"
  },
  "basket": {
    "position": [
      5,
      1
    ],
    "facing": "east"
  },
  "lamp": {
    "position": [
      -11,
      5
    ],
    "facing": "east"
  },
  "red door": {
    "position": [
      -5,
      0
    ],
    "facing": "east"
  },
  "green door": {
    "position": [
      2,
      0
    ],
    "facing": "east"
  },
  "agent": {
    "position": [
      5,
      1
    ],
    "facing": "south"
  }
}
📊 Cognitive Map JSONs
🤖 Predicted
{
  "origin": "agent",
  "objects": {
    "airplane": {
      "position": [
        2,
        4
      ],
      "facing": "+y"
    },
    "toilet": {
      "position": [
        -2,
        2
      ],
      "facing": "-x"
    },
    "fridge": {
      "position": [
        0,
        3
      ]
    }
  }
}
🎯 Ground Truth
{
  "toilet": {
    "position": [
      -2,
      2
    ],
    "facing": "east"
  },
  "airplane": {
    "position": [
      2,
      4
    ],
    "facing": "south"
  },
  "fridge": {
    "position": [
      0,
      4
    ],
    "facing": "east"
  }
}
📊 Cognitive Map JSONs
🤖 Predicted
{
  "origin": "agent",
  "objects": {
    "airplane": {
      "position": [
        2,
        4
      ],
      "facing": "+y"
    },
    "toilet": {
      "position": [
        -2,
        2
      ],
      "facing": "-x"
    },
    "fridge": {
      "position": [
        0,
        3
      ]
    }
  }
}
🎯 Ground Truth
{
  "airplane": {
    "position": [
      2,
      4
    ],
    "facing": "south"
  }
}
🧠 Cognitive Map Metrics
Global:
dir: 0.908
facing: 1
pos: 0.866
overall: 0.925
Global (Full):
dir: 0.908
facing: 1
pos: 0.866
overall: 0.925
Local:
dir: 0.833
facing: 1
pos: 0.918
overall: 0.917
Local (Newly):
dir: 1
facing: 1
pos: 1
overall: 1
Fog Probe:
(none)
📈 Turn Metrics
node_coverage: 1
edge_coverage: 0.537
is_action_fail: False
step: 13
action_counts:
move: 5
rotate: 8
return: 0
observe: 13
term: 0
forced_term: 0
query: 0
information_gain: 0.897
reward: -0.100
is_done: False
is_valid_action: True
Previous state
State before Turn 13
Current state
State at Turn 13

🔄 Turn 14

💬 Assistant Action
Actions: [Rotate(90), Observe()]
📊 Cognitive Map JSONs
🤖 Predicted
{
  "agent": {
    "position": [
      6,
      1
    ],
    "facing": "west"
  },
  "lego man": {
    "position": [
      -4,
      4
    ],
    "facing": "south"
  },
  "green door": {
    "position": [
      2,
      0
    ],
    "facing": "west"
  },
  "backpack": {
    "position": [
      -2,
      -3
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      2,
      -3
    ],
    "facing": "west"
  },
  "bike": {
    "position": [
      -5,
      1
    ],
    "facing": "south"
  },
  "red door": {
    "position": [
      -5,
      0
    ],
    "facing": "east"
  },
  "rubberduck": {
    "position": [
      -10,
      2
    ],
    "facing": "north"
  },
  "truck": {
    "position": [
      -7,
      2
    ],
    "facing": "east"
  },
  "lamp": {
    "position": [
      -8,
      4
    ]
  },
  "soda can": {
    "position": [
      -7,
      3
    ]
  },
  "toilet": {
    "position": [
      7,
      -1
    ],
    "facing": "west"
  },
  "fridge": {
    "position": [
      6,
      -2
    ]
  },
  "basket": {
    "position": [
      6,
      1
    ]
  },
  "airplane": {
    "position": [
      4,
      -2
    ],
    "facing": "north"
  }
}
🎯 Ground Truth (Observed)
{
  "bike": {
    "position": [
      -3,
      1
    ],
    "facing": "south"
  },
  "airplane": {
    "position": [
      3,
      -3
    ],
    "facing": "south"
  },
  "rubberduck": {
    "position": [
      -11,
      1
    ],
    "facing": "north"
  },
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  },
  "toilet": {
    "position": [
      7,
      -1
    ],
    "facing": "east"
  },
  "truck": {
    "position": [
      -7,
      1
    ],
    "facing": "east"
  },
  "backpack": {
    "position": [
      -1,
      -2
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      1,
      -2
    ],
    "facing": "west"
  },
  "fridge": {
    "position": [
      5,
      -3
    ],
    "facing": "east"
  },
  "soda can": {
    "position": [
      -7,
      3
    ],
    "facing": "east"
  },
  "basket": {
    "position": [
      5,
      1
    ],
    "facing": "east"
  },
  "lamp": {
    "position": [
      -11,
      5
    ],
    "facing": "east"
  },
  "red door": {
    "position": [
      -5,
      0
    ],
    "facing": "east"
  },
  "green door": {
    "position": [
      2,
      0
    ],
    "facing": "east"
  },
  "agent": {
    "position": [
      5,
      1
    ],
    "facing": "west"
  }
}
🎯 Ground Truth (Full)
{
  "bike": {
    "position": [
      -3,
      1
    ],
    "facing": "south"
  },
  "airplane": {
    "position": [
      3,
      -3
    ],
    "facing": "south"
  },
  "rubberduck": {
    "position": [
      -11,
      1
    ],
    "facing": "north"
  },
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  },
  "toilet": {
    "position": [
      7,
      -1
    ],
    "facing": "east"
  },
  "truck": {
    "position": [
      -7,
      1
    ],
    "facing": "east"
  },
  "backpack": {
    "position": [
      -1,
      -2
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      1,
      -2
    ],
    "facing": "west"
  },
  "fridge": {
    "position": [
      5,
      -3
    ],
    "facing": "east"
  },
  "soda can": {
    "position": [
      -7,
      3
    ],
    "facing": "east"
  },
  "basket": {
    "position": [
      5,
      1
    ],
    "facing": "east"
  },
  "lamp": {
    "position": [
      -11,
      5
    ],
    "facing": "east"
  },
  "red door": {
    "position": [
      -5,
      0
    ],
    "facing": "east"
  },
  "green door": {
    "position": [
      2,
      0
    ],
    "facing": "east"
  },
  "agent": {
    "position": [
      5,
      1
    ],
    "facing": "west"
  }
}
📊 Cognitive Map JSONs
🤖 Predicted
{
  "origin": "agent",
  "objects": {
    "green door": {
      "position": [
        -1,
        3
      ]
    }
  }
}
🎯 Ground Truth
{
  "green door": {
    "position": [
      -1,
      3
    ],
    "facing": "east"
  }
}
📊 Cognitive Map JSONs
🤖 Predicted
{
  "origin": "agent",
  "objects": {
    "green door": {
      "position": [
        -1,
        3
      ]
    }
  }
}
🎯 Ground Truth
(no data)
🧠 Cognitive Map Metrics
Global:
dir: 0.850
facing: 0.778
pos: 0.813
overall: 0.814
Global (Full):
dir: 0.850
facing: 0.778
pos: 0.813
overall: 0.814
Local:
dir: 1
facing: 1
pos: 1
overall: 1
Local (Newly):
(none)
Fog Probe:
(none)
📈 Turn Metrics
node_coverage: 1
edge_coverage: 0.537
is_action_fail: False
step: 14
action_counts:
move: 5
rotate: 9
return: 0
observe: 14
term: 0
forced_term: 0
query: 0
information_gain: 0.897
reward: -0.100
is_done: False
is_valid_action: True
Previous state
State before Turn 14
Current state
State at Turn 14

🔄 Turn 15

💬 Assistant Action
Actions: [Rotate(90), Observe()]
📊 Cognitive Map JSONs
🤖 Predicted
{
  "agent": {
    "position": [
      5,
      1
    ],
    "facing": "north"
  },
  "green door": {
    "position": [
      2,
      0
    ],
    "facing": "east"
  },
  "red door": {
    "position": [
      -5,
      0
    ],
    "facing": "west"
  },
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  },
  "backpack": {
    "position": [
      -2,
      -3
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      2,
      -3
    ],
    "facing": "west"
  },
  "bike": {
    "position": [
      -4,
      1
    ],
    "facing": "south"
  },
  "truck": {
    "position": [
      -7,
      2
    ],
    "facing": "east"
  },
  "soda can": {
    "position": [
      -7,
      3
    ]
  },
  "lamp": {
    "position": [
      -9,
      4
    ]
  },
  "rubberduck": {
    "position": [
      -10,
      2
    ],
    "facing": "north"
  },
  "basket": {
    "position": [
      5,
      1
    ]
  },
  "fridge": {
    "position": [
      5,
      -2
    ]
  },
  "toilet": {
    "position": [
      7,
      -2
    ],
    "facing": "south"
  },
  "airplane": {
    "position": [
      3,
      -2
    ],
    "facing": "south"
  }
}
🎯 Ground Truth (Observed)
{
  "bike": {
    "position": [
      -3,
      1
    ],
    "facing": "south"
  },
  "airplane": {
    "position": [
      3,
      -3
    ],
    "facing": "south"
  },
  "rubberduck": {
    "position": [
      -11,
      1
    ],
    "facing": "north"
  },
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  },
  "toilet": {
    "position": [
      7,
      -1
    ],
    "facing": "east"
  },
  "truck": {
    "position": [
      -7,
      1
    ],
    "facing": "east"
  },
  "backpack": {
    "position": [
      -1,
      -2
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      1,
      -2
    ],
    "facing": "west"
  },
  "fridge": {
    "position": [
      5,
      -3
    ],
    "facing": "east"
  },
  "soda can": {
    "position": [
      -7,
      3
    ],
    "facing": "east"
  },
  "basket": {
    "position": [
      5,
      1
    ],
    "facing": "east"
  },
  "lamp": {
    "position": [
      -11,
      5
    ],
    "facing": "east"
  },
  "red door": {
    "position": [
      -5,
      0
    ],
    "facing": "east"
  },
  "green door": {
    "position": [
      2,
      0
    ],
    "facing": "east"
  },
  "agent": {
    "position": [
      5,
      1
    ],
    "facing": "north"
  }
}
🎯 Ground Truth (Full)
{
  "bike": {
    "position": [
      -3,
      1
    ],
    "facing": "south"
  },
  "airplane": {
    "position": [
      3,
      -3
    ],
    "facing": "south"
  },
  "rubberduck": {
    "position": [
      -11,
      1
    ],
    "facing": "north"
  },
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  },
  "toilet": {
    "position": [
      7,
      -1
    ],
    "facing": "east"
  },
  "truck": {
    "position": [
      -7,
      1
    ],
    "facing": "east"
  },
  "backpack": {
    "position": [
      -1,
      -2
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      1,
      -2
    ],
    "facing": "west"
  },
  "fridge": {
    "position": [
      5,
      -3
    ],
    "facing": "east"
  },
  "soda can": {
    "position": [
      -7,
      3
    ],
    "facing": "east"
  },
  "basket": {
    "position": [
      5,
      1
    ],
    "facing": "east"
  },
  "lamp": {
    "position": [
      -11,
      5
    ],
    "facing": "east"
  },
  "red door": {
    "position": [
      -5,
      0
    ],
    "facing": "east"
  },
  "green door": {
    "position": [
      2,
      0
    ],
    "facing": "east"
  },
  "agent": {
    "position": [
      5,
      1
    ],
    "facing": "north"
  }
}
🧠 Cognitive Map Metrics
Global:
dir: 0.867
facing: 0.889
pos: 0.858
overall: 0.871
Global (Full):
dir: 0.867
facing: 0.889
pos: 0.858
overall: 0.871
Local:
(none)
Local (Newly):
(none)
Fog Probe:
(none)
📈 Turn Metrics
node_coverage: 1
edge_coverage: 0.537
is_action_fail: False
step: 15
action_counts:
move: 5
rotate: 10
return: 0
observe: 15
term: 0
forced_term: 0
query: 0
information_gain: 0.897
reward: -0.100
is_done: False
is_valid_action: True
Previous state
State before Turn 15
Current state
State at Turn 15

🔄 Turn 16

💬 Assistant Action
Actions: [Term()]
📊 Cognitive Map JSONs
🤖 Predicted
{
  "grid": "global",
  "frame": "origin at (0,0), north is +y",
  "agent": {
    "position": [
      5,
      1
    ],
    "facing": "north"
  },
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  },
  "green door": {
    "position": [
      2,
      0
    ],
    "facing": "west"
  },
  "backpack": {
    "position": [
      -2,
      -3
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      2,
      -3
    ],
    "facing": "west"
  },
  "bike": {
    "position": [
      -4,
      1
    ],
    "facing": "south"
  },
  "red door": {
    "position": [
      -5,
      0
    ],
    "facing": "east"
  },
  "rubberduck": {
    "position": [
      -10,
      2
    ],
    "facing": "north"
  },
  "truck": {
    "position": [
      -8,
      2
    ],
    "facing": "east"
  },
  "lamp": {
    "position": [
      -9,
      5
    ]
  },
  "soda can": {
    "position": [
      -8,
      3
    ]
  },
  "toilet": {
    "position": [
      8,
      -2
    ],
    "facing": "east"
  },
  "fridge": {
    "position": [
      5,
      -3
    ]
  },
  "basket": {
    "position": [
      5,
      1
    ]
  },
  "airplane": {
    "position": [
      2,
      -4
    ],
    "facing": "south"
  }
}
🎯 Ground Truth (Observed)
{
  "bike": {
    "position": [
      -3,
      1
    ],
    "facing": "south"
  },
  "airplane": {
    "position": [
      3,
      -3
    ],
    "facing": "south"
  },
  "rubberduck": {
    "position": [
      -11,
      1
    ],
    "facing": "north"
  },
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  },
  "toilet": {
    "position": [
      7,
      -1
    ],
    "facing": "east"
  },
  "truck": {
    "position": [
      -7,
      1
    ],
    "facing": "east"
  },
  "backpack": {
    "position": [
      -1,
      -2
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      1,
      -2
    ],
    "facing": "west"
  },
  "fridge": {
    "position": [
      5,
      -3
    ],
    "facing": "east"
  },
  "soda can": {
    "position": [
      -7,
      3
    ],
    "facing": "east"
  },
  "basket": {
    "position": [
      5,
      1
    ],
    "facing": "east"
  },
  "lamp": {
    "position": [
      -11,
      5
    ],
    "facing": "east"
  },
  "red door": {
    "position": [
      -5,
      0
    ],
    "facing": "east"
  },
  "green door": {
    "position": [
      2,
      0
    ],
    "facing": "east"
  },
  "agent": {
    "position": [
      5,
      1
    ],
    "facing": "north"
  }
}
🎯 Ground Truth (Full)
{
  "bike": {
    "position": [
      -3,
      1
    ],
    "facing": "south"
  },
  "airplane": {
    "position": [
      3,
      -3
    ],
    "facing": "south"
  },
  "rubberduck": {
    "position": [
      -11,
      1
    ],
    "facing": "north"
  },
  "lego man": {
    "position": [
      -3,
      3
    ],
    "facing": "south"
  },
  "toilet": {
    "position": [
      7,
      -1
    ],
    "facing": "east"
  },
  "truck": {
    "position": [
      -7,
      1
    ],
    "facing": "east"
  },
  "backpack": {
    "position": [
      -1,
      -2
    ],
    "facing": "south"
  },
  "shoppingcart": {
    "position": [
      1,
      -2
    ],
    "facing": "west"
  },
  "fridge": {
    "position": [
      5,
      -3
    ],
    "facing": "east"
  },
  "soda can": {
    "position": [
      -7,
      3
    ],
    "facing": "east"
  },
  "basket": {
    "position": [
      5,
      1
    ],
    "facing": "east"
  },
  "lamp": {
    "position": [
      -11,
      5
    ],
    "facing": "east"
  },
  "red door": {
    "position": [
      -5,
      0
    ],
    "facing": "east"
  },
  "green door": {
    "position": [
      2,
      0
    ],
    "facing": "east"
  },
  "agent": {
    "position": [
      5,
      1
    ],
    "facing": "north"
  }
}
🧠 Cognitive Map Metrics
Global:
dir: 0.858
facing: 1
pos: 0.850
overall: 0.903
Global (Full):
dir: 0.858
facing: 1
pos: 0.850
overall: 0.903
Local:
(none)
Local (Newly):
(none)
Fog Probe:
(none)
📈 Turn Metrics
node_coverage: 1
edge_coverage: 0.537
is_action_fail: False
step: 16
action_counts:
move: 5
rotate: 10
return: 0
observe: 15
term: 1
forced_term: 0
query: 0
information_gain: 0.897
reward: -0.100
is_done: True
is_valid_action: True
Previous state
State before Turn 16
Current state
State at Turn 16

🧭 False Belief Exploration

🔄 FB Turn 17

💬 Assistant Action
Actions: [Observe()]
📈 Turn Metrics
reward: -0.100
is_done: False
is_valid_action: True
🧭 False Belief Cogmap Metrics
Changed: lego man:
dir: None
pos: None
facing: 1
overall: None
Retention: lego man:
dir: None
pos: None
facing: 0
overall: None
Unchanged (all):
dir: 0.818
facing: 1
pos: 0.782
overall: 0.867
🧾 False Belief Object Sets
all_objects: ['airplane', 'backpack', 'basket', 'bike', 'fridge', 'green door', 'lamp', 'lego man', 'red door', 'rubberduck', 'shoppingcart', 'soda can', 'toilet', 'truck']
all_changed_objects: ['bike', 'lego man', 'shoppingcart', 'toilet']
newly_observed_changed (this turn): ['lego man']
unchanged_objects: ['airplane', 'backpack', 'basket', 'fridge', 'green door', 'lamp', 'red door', 'rubberduck', 'soda can', 'truck']
pred_changed_keys: ['lego man']
pred_unchanged_keys: ['airplane', 'backpack', 'basket', 'bike', 'fridge', 'green door', 'lamp', 'lego man', 'red door', 'rubberduck', 'shoppingcart', 'soda can', 'toilet', 'truck']
🧭 False Belief Info
step: 1
correctly_identified_changes: None
f1_overall: None
f1_position: None
f1_facing: None
action_cost: None
action_cost_after_seen: None
ground_truth_changes: [{'name': 'lego man', 'ori': True}, {'name': 'toilet', 'ori': True}, {'name': 'shoppingcart', 'pos': True}, {'name': 'bike', 'pos': True}]
reported_changes: []
newly_observed_changed_objects: ['lego man']
newly_observed_unchanged_objects: []
Current state
State at Turn 1

🔄 FB Turn 18

💬 Assistant Action
Actions: [Rotate(90), Observe()]
📈 Turn Metrics
reward: -0.100
is_done: False
is_valid_action: True
🧭 False Belief Cogmap Metrics
Unchanged (all):
dir: 0.803
facing: 0.600
pos: 0.822
overall: 0.742
🧾 False Belief Object Sets
all_objects: ['airplane', 'backpack', 'basket', 'bike', 'fridge', 'green door', 'lamp', 'lego man', 'red door', 'rubberduck', 'shoppingcart', 'soda can', 'toilet', 'truck']
all_changed_objects: ['bike', 'lego man', 'shoppingcart', 'toilet']
newly_observed_changed (this turn): []
unchanged_objects: ['airplane', 'backpack', 'basket', 'fridge', 'green door', 'lamp', 'red door', 'rubberduck', 'soda can', 'truck']
pred_changed_keys: []
pred_unchanged_keys: ['airplane', 'backpack', 'basket', 'bike', 'fridge', 'green door', 'lamp', 'lego man', 'red door', 'rubberduck', 'shoppingcart', 'soda can', 'toilet', 'truck']
🧭 False Belief Info
step: 2
correctly_identified_changes: None
f1_overall: None
f1_position: None
f1_facing: None
action_cost: None
action_cost_after_seen: None
ground_truth_changes: [{'name': 'lego man', 'ori': True}, {'name': 'toilet', 'ori': True}, {'name': 'shoppingcart', 'pos': True}, {'name': 'bike', 'pos': True}]
reported_changes: []
newly_observed_changed_objects: []
newly_observed_unchanged_objects: ['green door']
Previous state
State before Turn 2
Current state
State at Turn 2

🔄 FB Turn 19

💬 Assistant Action
Actions: [Rotate(90), Observe()]
📈 Turn Metrics
reward: -0.100
is_done: False
is_valid_action: True
🧭 False Belief Cogmap Metrics
Unchanged (all):
dir: 0.712
facing: 0.600
pos: 0.770
overall: 0.694
🧾 False Belief Object Sets
all_objects: ['airplane', 'backpack', 'basket', 'bike', 'fridge', 'green door', 'lamp', 'lego man', 'red door', 'rubberduck', 'shoppingcart', 'soda can', 'toilet', 'truck']
all_changed_objects: ['bike', 'lego man', 'shoppingcart', 'toilet']
newly_observed_changed (this turn): []
unchanged_objects: ['airplane', 'backpack', 'basket', 'fridge', 'green door', 'lamp', 'red door', 'rubberduck', 'soda can', 'truck']
pred_changed_keys: []
pred_unchanged_keys: ['airplane', 'backpack', 'basket', 'bike', 'fridge', 'green door', 'lamp', 'lego man', 'red door', 'rubberduck', 'shoppingcart', 'soda can', 'toilet', 'truck']
🧭 False Belief Info
step: 3
correctly_identified_changes: None
f1_overall: None
f1_position: None
f1_facing: None
action_cost: None
action_cost_after_seen: None
ground_truth_changes: [{'name': 'lego man', 'ori': True}, {'name': 'toilet', 'ori': True}, {'name': 'shoppingcart', 'pos': True}, {'name': 'bike', 'pos': True}]
reported_changes: []
newly_observed_changed_objects: []
newly_observed_unchanged_objects: ['backpack']
Previous state
State before Turn 3
Current state
State at Turn 3

🔄 FB Turn 20

💬 Assistant Action
Actions: [Rotate(90), Observe()]
📈 Turn Metrics
reward: -0.100
is_done: False
is_valid_action: True
🧭 False Belief Cogmap Metrics
Unchanged (all):
dir: 0.788
facing: 1
pos: 0.767
overall: 0.852
🧾 False Belief Object Sets
all_objects: ['airplane', 'backpack', 'basket', 'bike', 'fridge', 'green door', 'lamp', 'lego man', 'red door', 'rubberduck', 'shoppingcart', 'soda can', 'toilet', 'truck']
all_changed_objects: ['bike', 'lego man', 'shoppingcart', 'toilet']
newly_observed_changed (this turn): []
unchanged_objects: ['airplane', 'backpack', 'basket', 'fridge', 'green door', 'lamp', 'red door', 'rubberduck', 'soda can', 'truck']
pred_changed_keys: []
pred_unchanged_keys: ['airplane', 'backpack', 'basket', 'bike', 'fridge', 'green door', 'lamp', 'lego man', 'red door', 'rubberduck', 'shoppingcart', 'soda can', 'toilet', 'truck']
🧭 False Belief Info
step: 4
correctly_identified_changes: None
f1_overall: None
f1_position: None
f1_facing: None
action_cost: None
action_cost_after_seen: None
ground_truth_changes: [{'name': 'lego man', 'ori': True}, {'name': 'toilet', 'ori': True}, {'name': 'shoppingcart', 'pos': True}, {'name': 'bike', 'pos': True}]
reported_changes: []
newly_observed_changed_objects: []
newly_observed_unchanged_objects: ['red door']
Previous state
State before Turn 4
Current state
State at Turn 4

🔄 FB Turn 21

💬 Assistant Action
Actions: [JumpTo(red door), Observe()]
📈 Turn Metrics
reward: -0.100
is_done: False
is_valid_action: True
🧭 False Belief Cogmap Metrics
Changed: shoppingcart:
dir: None
pos: 0.475
facing: None
overall: None
Retention: shoppingcart:
dir: None
pos: 0.301
facing: None
overall: None
Unchanged (all):
dir: 0.515
facing: 0.400
pos: 0.602
overall: 0.506
🧾 False Belief Object Sets
all_objects: ['airplane', 'backpack', 'basket', 'bike', 'fridge', 'green door', 'lamp', 'lego man', 'red door', 'rubberduck', 'shoppingcart', 'soda can', 'toilet', 'truck']
all_changed_objects: ['bike', 'lego man', 'shoppingcart', 'toilet']
newly_observed_changed (this turn): ['shoppingcart']
unchanged_objects: ['airplane', 'backpack', 'basket', 'fridge', 'green door', 'lamp', 'red door', 'rubberduck', 'soda can', 'truck']
pred_changed_keys: ['shoppingcart']
pred_unchanged_keys: ['airplane', 'backpack', 'basket', 'bike', 'fridge', 'green door', 'lamp', 'lego man', 'red door', 'rubberduck', 'shoppingcart', 'soda can', 'toilet', 'truck']
🧭 False Belief Info
step: 5
correctly_identified_changes: None
f1_overall: None
f1_position: None
f1_facing: None
action_cost: None
action_cost_after_seen: None
ground_truth_changes: [{'name': 'lego man', 'ori': True}, {'name': 'toilet', 'ori': True}, {'name': 'shoppingcart', 'pos': True}, {'name': 'bike', 'pos': True}]
reported_changes: []
newly_observed_changed_objects: ['shoppingcart']
newly_observed_unchanged_objects: ['lamp', 'rubberduck', 'truck']
Previous state
State before Turn 5
Current state
State at Turn 5

🔄 FB Turn 22

💬 Assistant Action
Actions: [Rotate(180), Observe()]
📈 Turn Metrics
reward: -0.100
is_done: False
is_valid_action: True
🧭 False Belief Cogmap Metrics
Unchanged (all):
dir: 0.636
facing: 0.800
pos: 0.693
overall: 0.710
🧾 False Belief Object Sets
all_objects: ['airplane', 'backpack', 'basket', 'bike', 'fridge', 'green door', 'lamp', 'lego man', 'red door', 'rubberduck', 'shoppingcart', 'soda can', 'toilet', 'truck']
all_changed_objects: ['bike', 'lego man', 'shoppingcart', 'toilet']
newly_observed_changed (this turn): []
unchanged_objects: ['airplane', 'backpack', 'basket', 'fridge', 'green door', 'lamp', 'red door', 'rubberduck', 'soda can', 'truck']
pred_changed_keys: []
pred_unchanged_keys: ['airplane', 'backpack', 'basket', 'bike', 'fridge', 'green door', 'lamp', 'lego man', 'red door', 'rubberduck', 'shoppingcart', 'soda can', 'toilet', 'truck']
🧭 False Belief Info
step: 6
correctly_identified_changes: None
f1_overall: None
f1_position: None
f1_facing: None
action_cost: None
action_cost_after_seen: None
ground_truth_changes: [{'name': 'lego man', 'ori': True}, {'name': 'toilet', 'ori': True}, {'name': 'shoppingcart', 'pos': True}, {'name': 'bike', 'pos': True}]
reported_changes: []
newly_observed_changed_objects: []
newly_observed_unchanged_objects: []
Previous state
State before Turn 6
Current state
State at Turn 6

🔄 FB Turn 23

💬 Assistant Action
Actions: [JumpTo(green door), Observe()]
📈 Turn Metrics
reward: -0.100
is_done: False
is_valid_action: True
🧭 False Belief Cogmap Metrics
Changed: toilet:
dir: None
pos: None
facing: 1
overall: None
Retention: toilet:
dir: None
pos: None
facing: 0
overall: None
Changed: bike:
dir: None
pos: 0.862
facing: None
overall: None
Retention: bike:
dir: None
pos: 0.212
facing: None
overall: None
Unchanged (all):
dir: 0.788
facing: 1
pos: 0.753
overall: 0.847
🧾 False Belief Object Sets
all_objects: ['airplane', 'backpack', 'basket', 'bike', 'fridge', 'green door', 'lamp', 'lego man', 'red door', 'rubberduck', 'shoppingcart', 'soda can', 'toilet', 'truck']
all_changed_objects: ['bike', 'lego man', 'shoppingcart', 'toilet']
newly_observed_changed (this turn): ['bike', 'toilet']
unchanged_objects: ['airplane', 'backpack', 'basket', 'fridge', 'green door', 'lamp', 'red door', 'rubberduck', 'soda can', 'truck']
pred_changed_keys: ['bike', 'toilet']
pred_unchanged_keys: ['airplane', 'backpack', 'basket', 'bike', 'fridge', 'green door', 'lamp', 'lego man', 'red door', 'rubberduck', 'shoppingcart', 'soda can', 'toilet', 'truck']
🧭 False Belief Info
step: 7
correctly_identified_changes: None
f1_overall: None
f1_position: None
f1_facing: None
action_cost: None
action_cost_after_seen: None
ground_truth_changes: [{'name': 'lego man', 'ori': True}, {'name': 'toilet', 'ori': True}, {'name': 'shoppingcart', 'pos': True}, {'name': 'bike', 'pos': True}]
reported_changes: []
newly_observed_changed_objects: ['toilet', 'bike']
newly_observed_unchanged_objects: ['fridge', 'basket']
Previous state
State before Turn 7
Current state
State at Turn 7

🔄 FB Turn 24

💬 Assistant Action
Actions: [Term(changes="lego man: orientation, shoppingcart: position, bike: position, toilet: orientation")]
📈 Turn Metrics
reward: 1
is_done: True
is_valid_action: True
success: 1
🧭 False Belief Info
step: 8
correctly_identified_changes: 1
f1_overall: 1
f1_position: 1
f1_facing: 1
action_cost: 7
action_cost_after_seen: 0
ground_truth_changes: [{'name': 'lego man', 'ori': True}, {'name': 'toilet', 'ori': True}, {'name': 'shoppingcart', 'pos': True}, {'name': 'bike', 'pos': True}]
reported_changes: [{'name': 'lego man', 'ori': True}, {'name': 'shoppingcart', 'pos': True}, {'name': 'bike', 'pos': True}, {'name': 'toilet', 'ori': True}]
newly_observed_changed_objects: []
newly_observed_unchanged_objects: []
Previous state
State before Turn 8
Current state
State at Turn 8

📊 Evaluation (default)

📊 Task: DirectionEvaluationTask

Question 1 (ID: 9d13d2040c8f0e75)

✅ Evaluation Results
id: 9d13d2040c8f0e75
question: You return to your starting position and face north. From a Top-Down map, describe where basket is relative to toilet. Answer format: <cardinal direction>, <distance> Example: north-west, near
action: None
answer: north west, mid distance
task_type: DirectionEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: default
Evaluation state
Q1: DirectionEvaluationTask

Question 2 (ID: e2300d26c5895d2d)

✅ Evaluation Results
id: e2300d26c5895d2d
question: You return to your starting position and face north. From a Top-Down map, describe where rubberduck is relative to lego man. Answer format: <cardinal direction>, <distance> Example: north-west, near
action: None
answer: west, far
task_type: DirectionEvaluationTask
choices: []
kwargs:
(none)
score: 0.500
evaluation_mode: default
Evaluation state
Q2: DirectionEvaluationTask

Question 3 (ID: c2a21d8364e4dca4)

✅ Evaluation Results
id: c2a21d8364e4dca4
question: You return to your starting position and face north. From a Top-Down map, describe where shoppingcart is relative to airplane. Answer format: <cardinal direction>, <distance> Example: north-west, near
action: None
answer: north west, mid distance
task_type: DirectionEvaluationTask
choices: []
kwargs:
(none)
score: 0.500
evaluation_mode: default
Evaluation state
Q3: DirectionEvaluationTask

📊 Task: PovEvaluationTask

Question 1 (ID: 063960e28e84e1b8)

✅ Evaluation Results
id: 063960e28e84e1b8
question: Now you jump to rubberduck's direction, facing its direction. Describe where lamp is relative to you. Answer format: <ego direction>, <distance> Example: front-left, near
action: None
answer: front, mid distance
task_type: PovEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: default
Evaluation state
Q1: PovEvaluationTask

Question 2 (ID: dbbb7b93fba9d779)

✅ Evaluation Results
id: dbbb7b93fba9d779
question: Now you jump to lego man's direction, facing its direction. Describe where shoppingcart is relative to you. Answer format: <ego direction>, <distance> Example: front-left, near
action: None
answer: front-left, slightly far
task_type: PovEvaluationTask
choices: []
kwargs:
(none)
score: 0
evaluation_mode: default
Evaluation state
Q2: PovEvaluationTask

Question 3 (ID: b2e43f1e4b97a13e)

✅ Evaluation Results
id: b2e43f1e4b97a13e
question: Now you jump to shoppingcart's direction, facing its direction. Describe where bike is relative to you. Answer format: <ego direction>, <distance> Example: front-left, near
action: None
answer: front-right, slightly far
task_type: PovEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: default
Evaluation state
Q3: PovEvaluationTask

📊 Task: BackwardPovTextEvaluationTask

Question 1 (ID: 4022a79578213eb8)

✅ Evaluation Results
id: 4022a79578213eb8
question: Now you jump to an object's position, facing its direction. You observe that shoppingcart is front-left, slightly far, facing right; red door is front-right, mid distance, on right wall; backpack is front-slight-left, slightly far, facing forward. Which object are you standing at? Answer format: <object_name> Example: lamp
action: None
answer:
answer: lego man
final_pos: [9, 8]
final_ori: [0, -1]
final_observation: [{'name': 'shoppingcart', 'direction': 'front-left', 'distance': 'slightly far', 'orientation': 'facing right'}, {'name': 'red door', 'direction': 'front-right', 'distance': 'mid distance', 'orientation': 'on right wall'}, {'name': 'backpack', 'direction': 'front-slight-left', 'distance': 'slightly far', 'orientation': 'facing forward'}]
object_positions:
bike: [9.0, 6.0]
airplane: [15.0, 2.0]
rubberduck: [1.0, 6.0]
lego man: [9.0, 8.0]
toilet: [19.0, 4.0]
truck: [5.0, 6.0]
backpack: [11.0, 3.0]
shoppingcart: [13.0, 3.0]
fridge: [17.0, 2.0]
soda can: [5.0, 8.0]
basket: [17.0, 6.0]
lamp: [1.0, 10.0]
red door: [7.0, 5.0]
green door: [14.0, 5.0]
object_orientations:
bike: [0, -1]
airplane: [0, -1]
rubberduck: [0, 1]
lego man: [0, -1]
toilet: [1, 0]
truck: [1, 0]
backpack: [0, -1]
shoppingcart: [-1, 0]
red door: [1, 0]
green door: [1, 0]
room_id: 1
gate_info:
red door:
room_ids: [3, 1]
ori_by_room:
3: [-1, 0]
1: [1, 0]
green door:
room_ids: [1, 2]
ori_by_room:
1: [-1, 0]
2: [1, 0]
object_rooms:
bike: 1
airplane: 2
rubberduck: 3
lego man: 1
toilet: 2
truck: 3
backpack: 1
shoppingcart: 1
fridge: 2
soda can: 3
basket: 2
lamp: 3
red door: [3, 1]
green door: [1, 2]
task_type: BackwardPovTextEvaluationTask
choices: []
kwargs:
(none)
name_match: False
view_match: False
score: 0
evaluation_mode: default
Evaluation state
Q1: BackwardPovTextEvaluationTask

Question 2 (ID: d217d8691521707e)

✅ Evaluation Results
id: d217d8691521707e
question: Now you jump to an object's position, facing its direction. You observe that red door is front-right, mid distance, on front wall. Which object are you standing at? Answer format: <object_name> Example: lamp
action: None
answer:
answer: truck
final_pos: [5, 6]
final_ori: [1, 0]
final_observation: [{'name': 'red door', 'direction': 'front-right', 'distance': 'mid distance', 'orientation': 'on front wall'}]
object_positions:
bike: [9.0, 6.0]
airplane: [15.0, 2.0]
rubberduck: [1.0, 6.0]
lego man: [9.0, 8.0]
toilet: [19.0, 4.0]
truck: [5.0, 6.0]
backpack: [11.0, 3.0]
shoppingcart: [13.0, 3.0]
fridge: [17.0, 2.0]
soda can: [5.0, 8.0]
basket: [17.0, 6.0]
lamp: [1.0, 10.0]
red door: [7.0, 5.0]
green door: [14.0, 5.0]
object_orientations:
bike: [0, -1]
airplane: [0, -1]
rubberduck: [0, 1]
lego man: [0, -1]
toilet: [1, 0]
truck: [1, 0]
backpack: [0, -1]
shoppingcart: [-1, 0]
red door: [1, 0]
green door: [1, 0]
room_id: 3
gate_info:
red door:
room_ids: [3, 1]
ori_by_room:
3: [-1, 0]
1: [1, 0]
green door:
room_ids: [1, 2]
ori_by_room:
1: [-1, 0]
2: [1, 0]
object_rooms:
bike: 1
airplane: 2
rubberduck: 3
lego man: 1
toilet: 2
truck: 3
backpack: 1
shoppingcart: 1
fridge: 2
soda can: 3
basket: 2
lamp: 3
red door: [3, 1]
green door: [1, 2]
task_type: BackwardPovTextEvaluationTask
choices: []
kwargs:
(none)
name_match: True
view_match: True
score: 1
evaluation_mode: default
Evaluation state
Q2: BackwardPovTextEvaluationTask

Question 3 (ID: 7599c12dc388ac1a)

✅ Evaluation Results
id: 7599c12dc388ac1a
question: Now you jump to an object's position, facing its direction. You observe that lamp is front, mid distance. Which object are you standing at? Answer format: <object_name> Example: lamp
action: None
answer:
answer: rubberduck
final_pos: [1, 6]
final_ori: [0, 1]
final_observation: [{'name': 'lamp', 'direction': 'front', 'distance': 'mid distance', 'orientation': None}]
object_positions:
bike: [9.0, 6.0]
airplane: [15.0, 2.0]
rubberduck: [1.0, 6.0]
lego man: [9.0, 8.0]
toilet: [19.0, 4.0]
truck: [5.0, 6.0]
backpack: [11.0, 3.0]
shoppingcart: [13.0, 3.0]
fridge: [17.0, 2.0]
soda can: [5.0, 8.0]
basket: [17.0, 6.0]
lamp: [1.0, 10.0]
red door: [7.0, 5.0]
green door: [14.0, 5.0]
object_orientations:
bike: [0, -1]
airplane: [0, -1]
rubberduck: [0, 1]
lego man: [0, -1]
toilet: [1, 0]
truck: [1, 0]
backpack: [0, -1]
shoppingcart: [-1, 0]
red door: [1, 0]
green door: [1, 0]
room_id: 3
gate_info:
red door:
room_ids: [3, 1]
ori_by_room:
3: [-1, 0]
1: [1, 0]
green door:
room_ids: [1, 2]
ori_by_room:
1: [-1, 0]
2: [1, 0]
object_rooms:
bike: 1
airplane: 2
rubberduck: 3
lego man: 1
toilet: 2
truck: 3
backpack: 1
shoppingcart: 1
fridge: 2
soda can: 3
basket: 2
lamp: 3
red door: [3, 1]
green door: [1, 2]
task_type: BackwardPovTextEvaluationTask
choices: []
kwargs:
(none)
name_match: True
view_match: True
score: 1
evaluation_mode: default
Evaluation state
Q3: BackwardPovTextEvaluationTask

📊 Task: Action2ViewEvaluationTask

Question 1 (ID: e5146a0f564c765f)

✅ Evaluation Results
id: e5146a0f564c765f
question: You return to your starting position and face north. You will execute the following action sequence: 1. Rotate(90) 2. Jump to the object at front, near. 3. Jump to the object at front-right, slightly far. After executing the actions, what is the ego relation of toilet relative to you? Answer format: <ego direction>, <distance> Example: front, near
action: None
answer: front-left, mid distance
task_type: Action2ViewEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: default
Evaluation state
Q1: Action2ViewEvaluationTask

Question 2 (ID: 90f3cb1a03f64a70)

✅ Evaluation Results
id: 90f3cb1a03f64a70
question: You return to your starting position and face north. You will execute the following action sequence: 1. Rotate(90) 2. Jump to the object at front, near. 3. Jump to the object at front-right, slightly far. 4. Rotate(-180) After executing the actions, what is the ego relation of green door relative to you? Answer format: <ego direction>, <distance> Example: front, near
action: None
answer: front-right, slightly far
task_type: Action2ViewEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: default
Evaluation state
Q2: Action2ViewEvaluationTask

Question 3 (ID: 7b132df57af38164)

✅ Evaluation Results
id: 7b132df57af38164
question: You return to your starting position and face north. You will execute the following action sequence: 1. Jump to the object at front-left, slightly far. 2. Rotate(90) 3. Jump to the object at front-right, slightly far. 4. Rotate(90) After executing the actions, what is the ego relation of fridge relative to you? Answer format: <ego direction>, <distance> Example: front, near
action: None
answer: front-left, slightly far
task_type: Action2ViewEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: default
Evaluation state
Q3: Action2ViewEvaluationTask

📊 Task: View2ActionTextEvaluationTask

Question 1 (ID: a1b55f183063b5a2)

✅ Evaluation Results
id: a1b55f183063b5a2
question: You return to your starting position and face north. Then you have executed an action sequence and changed to a new location and facing direction. You observe the following: truck is at front, near, facing left; red door is at front-left, mid distance, on left wall What action sequence led to this final view? The action sequence must be valid and only contain move actions. Answer format: <sequence of move actions> Example: JumpTo(lamp), Rotate(90)
action: None
answer:
final_pos: [5, 8]
final_ori: [0, -1]
room_id: 3
init_pos: [12, 5]
init_ori: [0, 1]
object_positions:
bike: [9, 6]
airplane: [15, 2]
rubberduck: [1, 6]
lego man: [9, 8]
toilet: [19, 4]
truck: [5, 6]
backpack: [11, 3]
shoppingcart: [13, 3]
fridge: [17, 2]
soda can: [5, 8]
basket: [17, 6]
lamp: [1, 10]
red door: [7, 5]
green door: [14, 5]
object_orientations:
bike: [0, -1]
airplane: [0, -1]
rubberduck: [0, 1]
lego man: [0, -1]
toilet: [1, 0]
truck: [1, 0]
backpack: [0, -1]
shoppingcart: [-1, 0]
red door: [1, 0]
green door: [1, 0]
gate_info:
red door:
room_ids: [3, 1]
ori_by_room:
3: [-1, 0]
1: [1, 0]
green door:
room_ids: [1, 2]
ori_by_room:
1: [-1, 0]
2: [1, 0]
object_rooms:
bike: 1
airplane: 2
rubberduck: 3
lego man: 1
toilet: 2
truck: 3
backpack: 1
shoppingcart: 1
fridge: 2
soda can: 3
basket: 2
lamp: 3
red door: [3, 1]
green door: [1, 2]
minimal_plan: [['rotate', -90], ['jumpto', 'red door'], ['rotate', 90], ['jumpto', 'soda can'], ['rotate', 180]]
final_observation: [{'name': 'truck', 'direction': 'front', 'distance': 'near', 'orientation': 'facing left'}, {'name': 'red door', 'direction': 'front-left', 'distance': 'mid distance', 'orientation': 'on left wall'}]
task_type: View2ActionTextEvaluationTask
choices: []
kwargs:
(none)
pos_match: True
ori_match: True
final_pos: [5.0, 8.0]
final_ori: [0, -1]
visible_match: True
score: 1
evaluation_mode: default
Evaluation state
Q1: View2ActionTextEvaluationTask

Question 2 (ID: 514e1fc99b9dfec8)

✅ Evaluation Results
id: 514e1fc99b9dfec8
question: You return to your starting position and face north. Then you have executed an action sequence and changed to a new location and facing direction. You observe the following: basket is at front-right, slightly far; green door is at front-slight-left, mid distance, on left wall What action sequence led to this final view? The action sequence must be valid and only contain move actions. Answer format: <sequence of move actions> Example: JumpTo(lamp), Rotate(90)
action: None
answer:
final_pos: [15, 2]
final_ori: [0, 1]
room_id: 2
init_pos: [12, 5]
init_ori: [0, 1]
object_positions:
bike: [9, 6]
airplane: [15, 2]
rubberduck: [1, 6]
lego man: [9, 8]
toilet: [19, 4]
truck: [5, 6]
backpack: [11, 3]
shoppingcart: [13, 3]
fridge: [17, 2]
soda can: [5, 8]
basket: [17, 6]
lamp: [1, 10]
red door: [7, 5]
green door: [14, 5]
object_orientations:
bike: [0, -1]
airplane: [0, -1]
rubberduck: [0, 1]
lego man: [0, -1]
toilet: [1, 0]
truck: [1, 0]
backpack: [0, -1]
shoppingcart: [-1, 0]
red door: [1, 0]
green door: [1, 0]
gate_info:
red door:
room_ids: [3, 1]
ori_by_room:
3: [-1, 0]
1: [1, 0]
green door:
room_ids: [1, 2]
ori_by_room:
1: [-1, 0]
2: [1, 0]
object_rooms:
bike: 1
airplane: 2
rubberduck: 3
lego man: 1
toilet: 2
truck: 3
backpack: 1
shoppingcart: 1
fridge: 2
soda can: 3
basket: 2
lamp: 3
red door: [3, 1]
green door: [1, 2]
minimal_plan: [['rotate', 90], ['jumpto', 'green door'], ['rotate', 90], ['jumpto', 'airplane'], ['rotate', 180]]
final_observation: [{'name': 'basket', 'direction': 'front-right', 'distance': 'slightly far', 'orientation': None}, {'name': 'green door', 'direction': 'front-slight-left', 'distance': 'mid distance', 'orientation': 'on left wall'}]
task_type: View2ActionTextEvaluationTask
choices: []
kwargs:
(none)
error: invalid_format
score: 0
evaluation_mode: default
Evaluation state
Q2: View2ActionTextEvaluationTask

Question 3 (ID: bf8d3e056f97abbf)

✅ Evaluation Results
id: bf8d3e056f97abbf
question: You return to your starting position and face north. Then you have executed an action sequence and changed to a new location and facing direction. You observe the following: fridge is at front-right, mid distance What action sequence led to this final view? The action sequence must be valid and only contain move actions. Answer format: <sequence of move actions> Example: JumpTo(lamp), Rotate(90)
action: None
answer:
final_pos: [19, 4]
final_ori: [0, -1]
room_id: 2
init_pos: [12, 5]
init_ori: [0, 1]
object_positions:
bike: [9, 6]
airplane: [15, 2]
rubberduck: [1, 6]
lego man: [9, 8]
toilet: [19, 4]
truck: [5, 6]
backpack: [11, 3]
shoppingcart: [13, 3]
fridge: [17, 2]
soda can: [5, 8]
basket: [17, 6]
lamp: [1, 10]
red door: [7, 5]
green door: [14, 5]
object_orientations:
bike: [0, -1]
airplane: [0, -1]
rubberduck: [0, 1]
lego man: [0, -1]
toilet: [1, 0]
truck: [1, 0]
backpack: [0, -1]
shoppingcart: [-1, 0]
red door: [1, 0]
green door: [1, 0]
gate_info:
red door:
room_ids: [3, 1]
ori_by_room:
3: [-1, 0]
1: [1, 0]
green door:
room_ids: [1, 2]
ori_by_room:
1: [-1, 0]
2: [1, 0]
object_rooms:
bike: 1
airplane: 2
rubberduck: 3
lego man: 1
toilet: 2
truck: 3
backpack: 1
shoppingcart: 1
fridge: 2
soda can: 3
basket: 2
lamp: 3
red door: [3, 1]
green door: [1, 2]
minimal_plan: [['rotate', 90], ['jumpto', 'green door'], ['jumpto', 'toilet'], ['rotate', 90]]
final_observation: [{'name': 'fridge', 'direction': 'front-right', 'distance': 'mid distance', 'orientation': None}]
task_type: View2ActionTextEvaluationTask
choices: []
kwargs:
(none)
pos_match: True
ori_match: True
final_pos: [19.0, 4.0]
final_ori: [0, -1]
visible_match: True
score: 1
evaluation_mode: default
Evaluation state
Q3: View2ActionTextEvaluationTask

📊 Task: AlloMappingEvaluationTask

Question 1 (ID: 11808d88af254359)

✅ Evaluation Results
id: 11808d88af254359
question: Treat your starting position as the origin (0, 0) while facing north. Report allocentric coordinates using (x right/east, y up/north). Objects: basket, toilet, fridge. Answer format: (x0, y0); (x1, y1); ... in the same order. Example: (1, 0); (-2, 3); (0, -1)
action: None
answer: [[5, 1], [7, -1], [5, -3]]
task_type: AlloMappingEvaluationTask
choices: []
kwargs:
(none)
similarity: 0.924
threshold: 0.900
score: 0.924
evaluation_mode: default
Evaluation state
Q1: AlloMappingEvaluationTask

Question 2 (ID: 2b6ee8e5881229c1)

✅ Evaluation Results
id: 2b6ee8e5881229c1
question: Treat your starting position as the origin (0, 0) while facing north. Report allocentric coordinates using (x right/east, y up/north). Objects: rubberduck, lego man, basket, bike, airplane. Answer format: (x0, y0); (x1, y1); ... in the same order. Example: (1, 0); (-2, 3); (0, -1)
action: None
answer: [[-11, 1], [-3, 3], [5, 1], [-3, 1], [3, -3]]
task_type: AlloMappingEvaluationTask
choices: []
kwargs:
(none)
similarity: 0.942
threshold: 0.900
score: 0.942
evaluation_mode: default
Evaluation state
Q2: AlloMappingEvaluationTask

Question 3 (ID: 6a481cb096f39335)

✅ Evaluation Results
id: 6a481cb096f39335
question: Treat your starting position as the origin (0, 0) while facing north. Report allocentric coordinates using (x right/east, y up/north). Objects: backpack, basket, lamp, airplane, rubberduck. Answer format: (x0, y0); (x1, y1); ... in the same order. Example: (1, 0); (-2, 3); (0, -1)
action: None
answer: [[-1, -2], [5, 1], [-11, 5], [3, -3], [-11, 1]]
task_type: AlloMappingEvaluationTask
choices: []
kwargs:
(none)
similarity: 0.919
threshold: 0.900
score: 0.919
evaluation_mode: default
Evaluation state
Q3: AlloMappingEvaluationTask

📊 Task: RotEvaluationTask

Question 1 (ID: 37b10f92ed55a226)

✅ Evaluation Results
id: 37b10f92ed55a226
question: You return to your starting position and face north. You will perform a full 360-degree rotation by continuously turning clockwise in place. Assume all walls are removed (you can see through walls), so every object is visible. Focus on this set of objects: basket, fridge, shoppingcart. List them in the exact order they appear directly ahead while you rotate. If two objects share a bearing, place the nearer one first. Answer format: <object_name1>, <object_name2>, ... Example: mug, sofa, plant
action: None
answer: ['basket', 'fridge', 'shoppingcart']
task_type: RotEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: default
Evaluation state
Q1: RotEvaluationTask

Question 2 (ID: 4b2dc57e77686a8b)

✅ Evaluation Results
id: 4b2dc57e77686a8b
question: You return to your starting position and face north. You will perform a full 360-degree rotation by continuously turning counterclockwise in place. Assume all walls are removed (you can see through walls), so every object is visible. Focus on this set of objects: backpack, lego man, rubberduck. List them in the exact order they appear directly ahead while you rotate. If two objects share a bearing, place the nearer one first. Answer format: <object_name1>, <object_name2>, ... Example: mug, sofa, plant
action: None
answer: ['lego man', 'rubberduck', 'backpack']
task_type: RotEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: default
Evaluation state
Q2: RotEvaluationTask

Question 3 (ID: 3af660d5e0016740)

✅ Evaluation Results
id: 3af660d5e0016740
question: You return to your starting position and face north. You will perform a full 360-degree rotation by continuously turning clockwise in place. Assume all walls are removed (you can see through walls), so every object is visible. Focus on this set of objects: airplane, backpack, lamp, toilet. List them in the exact order they appear directly ahead while you rotate. If two objects share a bearing, place the nearer one first. Answer format: <object_name1>, <object_name2>, ... Example: mug, sofa, plant
action: None
answer: ['toilet', 'airplane', 'backpack', 'lamp']
task_type: RotEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: default
Evaluation state
Q3: RotEvaluationTask

📊 Task: RotDualEvaluationTask

Question 1 (ID: 03400bfc656221c4)

✅ Evaluation Results
id: 03400bfc656221c4
question: You return to your starting position and face north. You performed a complete 360° rotation in place. Assume all walls are removed (you can see through walls), so every object is visible. During the rotation, these objects appeared directly in front of you in this order: basket, fridge, shoppingcart Based on this sequence, in which direction did you rotate? Answer format: clockwise or counterclockwise Example: clockwise
action: None
answer: clockwise
task_type: RotDualEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: default
Evaluation state
Q1: RotDualEvaluationTask

Question 2 (ID: 6d8e9eae3f521541)

✅ Evaluation Results
id: 6d8e9eae3f521541
question: You return to your starting position and face north. You performed a complete 360° rotation in place. Assume all walls are removed (you can see through walls), so every object is visible. During the rotation, these objects appeared directly in front of you in this order: lego man, rubberduck, backpack Based on this sequence, in which direction did you rotate? Answer format: clockwise or counterclockwise Example: clockwise
action: None
answer: counterclockwise
task_type: RotDualEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: default
Evaluation state
Q2: RotDualEvaluationTask

Question 3 (ID: a0d83aabc0d2bbd6)

✅ Evaluation Results
id: a0d83aabc0d2bbd6
question: You return to your starting position and face north. You performed a complete 360° rotation in place. Assume all walls are removed (you can see through walls), so every object is visible. During the rotation, these objects appeared directly in front of you in this order: toilet, airplane, backpack, lamp Based on this sequence, in which direction did you rotate? Answer format: clockwise or counterclockwise Example: clockwise
action: None
answer: clockwise
task_type: RotDualEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: default
Evaluation state
Q3: RotDualEvaluationTask

📊 Task: Location2ViewEvaluationTask

Question 1 (ID: 3af268195886a8df)

✅ Evaluation Results
id: 3af268195886a8df
question: Treat the green door as the new 'origin' (0, 0). You move to (2, -2) and face east. What is the egocentric relation of toilet? Answer format: <direction>, <distance> Example: front, near
action: None
answer: front-slight-left, mid distance
task_type: Action2LocationEvaluationTask
choices: []
kwargs:
(none)
score: 0.500
evaluation_mode: default
Evaluation state
Q1: Location2ViewEvaluationTask

Question 2 (ID: 56728a7c1f765ce3)

✅ Evaluation Results
id: 56728a7c1f765ce3
question: Treat the green door as the new 'origin' (0, 0). You move to (2, -2) and face north. What is the egocentric relation of basket? Answer format: <direction>, <distance> Example: front, near
action: None
answer: front-slight-right, mid distance
task_type: Action2LocationEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: default
Evaluation state
Q2: Location2ViewEvaluationTask

Question 3 (ID: 843e8e1b199cd022)

✅ Evaluation Results
id: 843e8e1b199cd022
question: Treat the red door as the new 'origin' (0, 0). You move to (-3, 4) and face south. What is the egocentric relation of truck? Answer format: <direction>, <distance> Example: front, near
action: None
answer: front-slight-left, mid distance
task_type: Action2LocationEvaluationTask
choices: []
kwargs:
(none)
score: 0
evaluation_mode: default
Evaluation state
Q3: Location2ViewEvaluationTask

📊 Task: View2LocationTextEvaluationTask

Question 1 (ID: 0627604bb51f9403)

✅ Evaluation Results
id: 0627604bb51f9403
question: You move to a new location and face east. You observe: toilet is at front-slight-left, mid distance, facing forward; fridge is at front-right, near Treat the green door as the new 'origin' (0, 0). What is your new 2D coordinate (x, y)? Answer format: (x, y) Example: (2, -1)
action: None
answer:
coord: [2, -2]
final_pos: [16, 3]
final_ori: [1, 0]
room_id: 2
object_positions:
bike: [9, 6]
airplane: [15, 2]
rubberduck: [1, 6]
lego man: [9, 8]
toilet: [19, 4]
truck: [5, 6]
backpack: [11, 3]
shoppingcart: [13, 3]
fridge: [17, 2]
soda can: [5, 8]
basket: [17, 6]
lamp: [1, 10]
red door: [7, 5]
green door: [14, 5]
object_orientations:
bike: [0, -1]
airplane: [0, -1]
rubberduck: [0, 1]
lego man: [0, -1]
toilet: [1, 0]
truck: [1, 0]
backpack: [0, -1]
shoppingcart: [-1, 0]
red door: [1, 0]
green door: [1, 0]
final_observation: [{'name': 'toilet', 'direction': 'front-slight-left', 'distance': 'mid distance', 'orientation': 'facing forward'}, {'name': 'fridge', 'direction': 'front-right', 'distance': 'near', 'orientation': None}]
gate_info:
red door:
room_ids: [3, 1]
ori_by_room:
3: [-1, 0]
1: [1, 0]
green door:
room_ids: [1, 2]
ori_by_room:
1: [-1, 0]
2: [1, 0]
object_rooms:
bike: 1
airplane: 2
rubberduck: 3
lego man: 1
toilet: 2
truck: 3
backpack: 1
shoppingcart: 1
fridge: 2
soda can: 3
basket: 2
lamp: 3
red door: [3, 1]
green door: [1, 2]
task_type: Location2ActionTextEvaluationTask
choices: []
kwargs:
(none)
similarity: 1
raw_score: 1
score: 1
evaluation_mode: default
Evaluation state
Q1: View2LocationTextEvaluationTask

Question 2 (ID: 9d51c9ea1c03280c)

✅ Evaluation Results
id: 9d51c9ea1c03280c
question: You move to a new location and face north. You observe: basket is at front-slight-right, mid distance; green door is at front-left, mid distance, on left wall Treat the green door as the new 'origin' (0, 0). What is your new 2D coordinate (x, y)? Answer format: (x, y) Example: (2, -1)
action: None
answer:
coord: [2, -2]
final_pos: [16, 3]
final_ori: [0, 1]
room_id: 2
object_positions:
bike: [9, 6]
airplane: [15, 2]
rubberduck: [1, 6]
lego man: [9, 8]
toilet: [19, 4]
truck: [5, 6]
backpack: [11, 3]
shoppingcart: [13, 3]
fridge: [17, 2]
soda can: [5, 8]
basket: [17, 6]
lamp: [1, 10]
red door: [7, 5]
green door: [14, 5]
object_orientations:
bike: [0, -1]
airplane: [0, -1]
rubberduck: [0, 1]
lego man: [0, -1]
toilet: [1, 0]
truck: [1, 0]
backpack: [0, -1]
shoppingcart: [-1, 0]
red door: [1, 0]
green door: [1, 0]
final_observation: [{'name': 'basket', 'direction': 'front-slight-right', 'distance': 'mid distance', 'orientation': None}, {'name': 'green door', 'direction': 'front-left', 'distance': 'mid distance', 'orientation': 'on left wall'}]
gate_info:
red door:
room_ids: [3, 1]
ori_by_room:
3: [-1, 0]
1: [1, 0]
green door:
room_ids: [1, 2]
ori_by_room:
1: [-1, 0]
2: [1, 0]
object_rooms:
bike: 1
airplane: 2
rubberduck: 3
lego man: 1
toilet: 2
truck: 3
backpack: 1
shoppingcart: 1
fridge: 2
soda can: 3
basket: 2
lamp: 3
red door: [3, 1]
green door: [1, 2]
task_type: Location2ActionTextEvaluationTask
choices: []
kwargs:
(none)
similarity: 1
raw_score: 1
score: 1
evaluation_mode: default
Evaluation state
Q2: View2LocationTextEvaluationTask

Question 3 (ID: 231594f7927a5d40)

✅ Evaluation Results
id: 231594f7927a5d40
question: You move to a new location and face south. You observe: truck is at front-slight-left, mid distance, facing left; rubberduck is at front-right, slightly far, facing backward; red door is at front-left, slightly far, on left wall Treat the red door as the new 'origin' (0, 0). What is your new 2D coordinate (x, y)? Answer format: (x, y) Example: (2, -1)
action: None
answer:
coord: [-3, 4]
final_pos: [4, 9]
final_ori: [0, -1]
room_id: 3
object_positions:
bike: [9, 6]
airplane: [15, 2]
rubberduck: [1, 6]
lego man: [9, 8]
toilet: [19, 4]
truck: [5, 6]
backpack: [11, 3]
shoppingcart: [13, 3]
fridge: [17, 2]
soda can: [5, 8]
basket: [17, 6]
lamp: [1, 10]
red door: [7, 5]
green door: [14, 5]
object_orientations:
bike: [0, -1]
airplane: [0, -1]
rubberduck: [0, 1]
lego man: [0, -1]
toilet: [1, 0]
truck: [1, 0]
backpack: [0, -1]
shoppingcart: [-1, 0]
red door: [1, 0]
green door: [1, 0]
final_observation: [{'name': 'truck', 'direction': 'front-slight-left', 'distance': 'mid distance', 'orientation': 'facing left'}, {'name': 'rubberduck', 'direction': 'front-right', 'distance': 'slightly far', 'orientation': 'facing backward'}, {'name': 'red door', 'direction': 'front-left', 'distance': 'slightly far', 'orientation': 'on left wall'}]
gate_info:
red door:
room_ids: [3, 1]
ori_by_room:
3: [-1, 0]
1: [1, 0]
green door:
room_ids: [1, 2]
ori_by_room:
1: [-1, 0]
2: [1, 0]
object_rooms:
bike: 1
airplane: 2
rubberduck: 3
lego man: 1
toilet: 2
truck: 3
backpack: 1
shoppingcart: 1
fridge: 2
soda can: 3
basket: 2
lamp: 3
red door: [3, 1]
green door: [1, 2]
task_type: Location2ActionTextEvaluationTask
choices: []
kwargs:
(none)
similarity: 0.819
raw_score: 0.819
score: 0.819
evaluation_mode: default
Evaluation state
Q3: View2LocationTextEvaluationTask

📊 Evaluation (prompt_cogmap)

📊 Task: DirectionEvaluationTask

Question 1 (ID: 9d13d2040c8f0e75)

✅ Evaluation Results
id: 9d13d2040c8f0e75
question: You return to your starting position and face north. From a Top-Down map, describe where basket is relative to toilet. Answer format: <cardinal direction>, <distance> Example: north-west, near
action: None
answer: north west, mid distance
task_type: DirectionEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: prompt_cogmap
Evaluation state
Q1: DirectionEvaluationTask

Question 2 (ID: e2300d26c5895d2d)

✅ Evaluation Results
id: e2300d26c5895d2d
question: You return to your starting position and face north. From a Top-Down map, describe where rubberduck is relative to lego man. Answer format: <cardinal direction>, <distance> Example: north-west, near
action: None
answer: west, far
task_type: DirectionEvaluationTask
choices: []
kwargs:
(none)
score: 0.500
evaluation_mode: prompt_cogmap
Evaluation state
Q2: DirectionEvaluationTask

Question 3 (ID: c2a21d8364e4dca4)

✅ Evaluation Results
id: c2a21d8364e4dca4
question: You return to your starting position and face north. From a Top-Down map, describe where shoppingcart is relative to airplane. Answer format: <cardinal direction>, <distance> Example: north-west, near
action: None
answer: north west, mid distance
task_type: DirectionEvaluationTask
choices: []
kwargs:
(none)
score: 0
evaluation_mode: prompt_cogmap
Evaluation state
Q3: DirectionEvaluationTask

📊 Task: PovEvaluationTask

Question 1 (ID: 063960e28e84e1b8)

✅ Evaluation Results
id: 063960e28e84e1b8
question: Now you jump to rubberduck's direction, facing its direction. Describe where lamp is relative to you. Answer format: <ego direction>, <distance> Example: front-left, near
action: None
answer: front, mid distance
task_type: PovEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: prompt_cogmap
Evaluation state
Q1: PovEvaluationTask

Question 2 (ID: dbbb7b93fba9d779)

✅ Evaluation Results
id: dbbb7b93fba9d779
question: Now you jump to lego man's direction, facing its direction. Describe where shoppingcart is relative to you. Answer format: <ego direction>, <distance> Example: front-left, near
action: None
answer: front-left, slightly far
task_type: PovEvaluationTask
choices: []
kwargs:
(none)
score: 0.500
evaluation_mode: prompt_cogmap
Evaluation state
Q2: PovEvaluationTask

Question 3 (ID: b2e43f1e4b97a13e)

✅ Evaluation Results
id: b2e43f1e4b97a13e
question: Now you jump to shoppingcart's direction, facing its direction. Describe where bike is relative to you. Answer format: <ego direction>, <distance> Example: front-left, near
action: None
answer: front-right, slightly far
task_type: PovEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: prompt_cogmap
Evaluation state
Q3: PovEvaluationTask

📊 Task: BackwardPovTextEvaluationTask

Question 1 (ID: 4022a79578213eb8)

✅ Evaluation Results
id: 4022a79578213eb8
question: Now you jump to an object's position, facing its direction. You observe that shoppingcart is front-left, slightly far, facing right; red door is front-right, mid distance, on right wall; backpack is front-slight-left, slightly far, facing forward. Which object are you standing at? Answer format: <object_name> Example: lamp
action: None
answer:
answer: lego man
final_pos: [9, 8]
final_ori: [0, -1]
final_observation: [{'name': 'shoppingcart', 'direction': 'front-left', 'distance': 'slightly far', 'orientation': 'facing right'}, {'name': 'red door', 'direction': 'front-right', 'distance': 'mid distance', 'orientation': 'on right wall'}, {'name': 'backpack', 'direction': 'front-slight-left', 'distance': 'slightly far', 'orientation': 'facing forward'}]
object_positions:
bike: [9.0, 6.0]
airplane: [15.0, 2.0]
rubberduck: [1.0, 6.0]
lego man: [9.0, 8.0]
toilet: [19.0, 4.0]
truck: [5.0, 6.0]
backpack: [11.0, 3.0]
shoppingcart: [13.0, 3.0]
fridge: [17.0, 2.0]
soda can: [5.0, 8.0]
basket: [17.0, 6.0]
lamp: [1.0, 10.0]
red door: [7.0, 5.0]
green door: [14.0, 5.0]
object_orientations:
bike: [0, -1]
airplane: [0, -1]
rubberduck: [0, 1]
lego man: [0, -1]
toilet: [1, 0]
truck: [1, 0]
backpack: [0, -1]
shoppingcart: [-1, 0]
red door: [1, 0]
green door: [1, 0]
room_id: 1
gate_info:
red door:
room_ids: [3, 1]
ori_by_room:
3: [-1, 0]
1: [1, 0]
green door:
room_ids: [1, 2]
ori_by_room:
1: [-1, 0]
2: [1, 0]
object_rooms:
bike: 1
airplane: 2
rubberduck: 3
lego man: 1
toilet: 2
truck: 3
backpack: 1
shoppingcart: 1
fridge: 2
soda can: 3
basket: 2
lamp: 3
red door: [3, 1]
green door: [1, 2]
task_type: BackwardPovTextEvaluationTask
choices: []
kwargs:
(none)
name_match: False
view_match: False
score: 0
evaluation_mode: prompt_cogmap
Evaluation state
Q1: BackwardPovTextEvaluationTask

Question 2 (ID: d217d8691521707e)

✅ Evaluation Results
id: d217d8691521707e
question: Now you jump to an object's position, facing its direction. You observe that red door is front-right, mid distance, on front wall. Which object are you standing at? Answer format: <object_name> Example: lamp
action: None
answer:
answer: truck
final_pos: [5, 6]
final_ori: [1, 0]
final_observation: [{'name': 'red door', 'direction': 'front-right', 'distance': 'mid distance', 'orientation': 'on front wall'}]
object_positions:
bike: [9.0, 6.0]
airplane: [15.0, 2.0]
rubberduck: [1.0, 6.0]
lego man: [9.0, 8.0]
toilet: [19.0, 4.0]
truck: [5.0, 6.0]
backpack: [11.0, 3.0]
shoppingcart: [13.0, 3.0]
fridge: [17.0, 2.0]
soda can: [5.0, 8.0]
basket: [17.0, 6.0]
lamp: [1.0, 10.0]
red door: [7.0, 5.0]
green door: [14.0, 5.0]
object_orientations:
bike: [0, -1]
airplane: [0, -1]
rubberduck: [0, 1]
lego man: [0, -1]
toilet: [1, 0]
truck: [1, 0]
backpack: [0, -1]
shoppingcart: [-1, 0]
red door: [1, 0]
green door: [1, 0]
room_id: 3
gate_info:
red door:
room_ids: [3, 1]
ori_by_room:
3: [-1, 0]
1: [1, 0]
green door:
room_ids: [1, 2]
ori_by_room:
1: [-1, 0]
2: [1, 0]
object_rooms:
bike: 1
airplane: 2
rubberduck: 3
lego man: 1
toilet: 2
truck: 3
backpack: 1
shoppingcart: 1
fridge: 2
soda can: 3
basket: 2
lamp: 3
red door: [3, 1]
green door: [1, 2]
task_type: BackwardPovTextEvaluationTask
choices: []
kwargs:
(none)
name_match: True
view_match: True
score: 1
evaluation_mode: prompt_cogmap
Evaluation state
Q2: BackwardPovTextEvaluationTask

Question 3 (ID: 7599c12dc388ac1a)

✅ Evaluation Results
id: 7599c12dc388ac1a
question: Now you jump to an object's position, facing its direction. You observe that lamp is front, mid distance. Which object are you standing at? Answer format: <object_name> Example: lamp
action: None
answer:
answer: rubberduck
final_pos: [1, 6]
final_ori: [0, 1]
final_observation: [{'name': 'lamp', 'direction': 'front', 'distance': 'mid distance', 'orientation': None}]
object_positions:
bike: [9.0, 6.0]
airplane: [15.0, 2.0]
rubberduck: [1.0, 6.0]
lego man: [9.0, 8.0]
toilet: [19.0, 4.0]
truck: [5.0, 6.0]
backpack: [11.0, 3.0]
shoppingcart: [13.0, 3.0]
fridge: [17.0, 2.0]
soda can: [5.0, 8.0]
basket: [17.0, 6.0]
lamp: [1.0, 10.0]
red door: [7.0, 5.0]
green door: [14.0, 5.0]
object_orientations:
bike: [0, -1]
airplane: [0, -1]
rubberduck: [0, 1]
lego man: [0, -1]
toilet: [1, 0]
truck: [1, 0]
backpack: [0, -1]
shoppingcart: [-1, 0]
red door: [1, 0]
green door: [1, 0]
room_id: 3
gate_info:
red door:
room_ids: [3, 1]
ori_by_room:
3: [-1, 0]
1: [1, 0]
green door:
room_ids: [1, 2]
ori_by_room:
1: [-1, 0]
2: [1, 0]
object_rooms:
bike: 1
airplane: 2
rubberduck: 3
lego man: 1
toilet: 2
truck: 3
backpack: 1
shoppingcart: 1
fridge: 2
soda can: 3
basket: 2
lamp: 3
red door: [3, 1]
green door: [1, 2]
task_type: BackwardPovTextEvaluationTask
choices: []
kwargs:
(none)
name_match: True
view_match: True
score: 1
evaluation_mode: prompt_cogmap
Evaluation state
Q3: BackwardPovTextEvaluationTask

📊 Task: Action2ViewEvaluationTask

Question 1 (ID: e5146a0f564c765f)

✅ Evaluation Results
id: e5146a0f564c765f
question: You return to your starting position and face north. You will execute the following action sequence: 1. Rotate(90) 2. Jump to the object at front, near. 3. Jump to the object at front-right, slightly far. After executing the actions, what is the ego relation of toilet relative to you? Answer format: <ego direction>, <distance> Example: front, near
action: None
answer: front-left, mid distance
task_type: Action2ViewEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: prompt_cogmap
Evaluation state
Q1: Action2ViewEvaluationTask

Question 2 (ID: 90f3cb1a03f64a70)

✅ Evaluation Results
id: 90f3cb1a03f64a70
question: You return to your starting position and face north. You will execute the following action sequence: 1. Rotate(90) 2. Jump to the object at front, near. 3. Jump to the object at front-right, slightly far. 4. Rotate(-180) After executing the actions, what is the ego relation of green door relative to you? Answer format: <ego direction>, <distance> Example: front, near
action: None
answer: front-right, slightly far
task_type: Action2ViewEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: prompt_cogmap
Evaluation state
Q2: Action2ViewEvaluationTask

Question 3 (ID: 7b132df57af38164)

✅ Evaluation Results
id: 7b132df57af38164
question: You return to your starting position and face north. You will execute the following action sequence: 1. Jump to the object at front-left, slightly far. 2. Rotate(90) 3. Jump to the object at front-right, slightly far. 4. Rotate(90) After executing the actions, what is the ego relation of fridge relative to you? Answer format: <ego direction>, <distance> Example: front, near
action: None
answer: front-left, slightly far
task_type: Action2ViewEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: prompt_cogmap
Evaluation state
Q3: Action2ViewEvaluationTask

📊 Task: View2ActionTextEvaluationTask

Question 1 (ID: a1b55f183063b5a2)

✅ Evaluation Results
id: a1b55f183063b5a2
question: You return to your starting position and face north. Then you have executed an action sequence and changed to a new location and facing direction. You observe the following: truck is at front, near, facing left; red door is at front-left, mid distance, on left wall What action sequence led to this final view? The action sequence must be valid and only contain move actions. Answer format: <sequence of move actions> Example: JumpTo(lamp), Rotate(90)
action: None
answer:
final_pos: [5, 8]
final_ori: [0, -1]
room_id: 3
init_pos: [12, 5]
init_ori: [0, 1]
object_positions:
bike: [9, 6]
airplane: [15, 2]
rubberduck: [1, 6]
lego man: [9, 8]
toilet: [19, 4]
truck: [5, 6]
backpack: [11, 3]
shoppingcart: [13, 3]
fridge: [17, 2]
soda can: [5, 8]
basket: [17, 6]
lamp: [1, 10]
red door: [7, 5]
green door: [14, 5]
object_orientations:
bike: [0, -1]
airplane: [0, -1]
rubberduck: [0, 1]
lego man: [0, -1]
toilet: [1, 0]
truck: [1, 0]
backpack: [0, -1]
shoppingcart: [-1, 0]
red door: [1, 0]
green door: [1, 0]
gate_info:
red door:
room_ids: [3, 1]
ori_by_room:
3: [-1, 0]
1: [1, 0]
green door:
room_ids: [1, 2]
ori_by_room:
1: [-1, 0]
2: [1, 0]
object_rooms:
bike: 1
airplane: 2
rubberduck: 3
lego man: 1
toilet: 2
truck: 3
backpack: 1
shoppingcart: 1
fridge: 2
soda can: 3
basket: 2
lamp: 3
red door: [3, 1]
green door: [1, 2]
minimal_plan: [['rotate', -90], ['jumpto', 'red door'], ['rotate', 90], ['jumpto', 'soda can'], ['rotate', 180]]
final_observation: [{'name': 'truck', 'direction': 'front', 'distance': 'near', 'orientation': 'facing left'}, {'name': 'red door', 'direction': 'front-left', 'distance': 'mid distance', 'orientation': 'on left wall'}]
task_type: View2ActionTextEvaluationTask
choices: []
kwargs:
(none)
pos_match: True
ori_match: True
final_pos: [5.0, 8.0]
final_ori: [0, -1]
visible_match: True
score: 1
evaluation_mode: prompt_cogmap
Evaluation state
Q1: View2ActionTextEvaluationTask

Question 2 (ID: 514e1fc99b9dfec8)

✅ Evaluation Results
id: 514e1fc99b9dfec8
question: You return to your starting position and face north. Then you have executed an action sequence and changed to a new location and facing direction. You observe the following: basket is at front-right, slightly far; green door is at front-slight-left, mid distance, on left wall What action sequence led to this final view? The action sequence must be valid and only contain move actions. Answer format: <sequence of move actions> Example: JumpTo(lamp), Rotate(90)
action: None
answer:
final_pos: [15, 2]
final_ori: [0, 1]
room_id: 2
init_pos: [12, 5]
init_ori: [0, 1]
object_positions:
bike: [9, 6]
airplane: [15, 2]
rubberduck: [1, 6]
lego man: [9, 8]
toilet: [19, 4]
truck: [5, 6]
backpack: [11, 3]
shoppingcart: [13, 3]
fridge: [17, 2]
soda can: [5, 8]
basket: [17, 6]
lamp: [1, 10]
red door: [7, 5]
green door: [14, 5]
object_orientations:
bike: [0, -1]
airplane: [0, -1]
rubberduck: [0, 1]
lego man: [0, -1]
toilet: [1, 0]
truck: [1, 0]
backpack: [0, -1]
shoppingcart: [-1, 0]
red door: [1, 0]
green door: [1, 0]
gate_info:
red door:
room_ids: [3, 1]
ori_by_room:
3: [-1, 0]
1: [1, 0]
green door:
room_ids: [1, 2]
ori_by_room:
1: [-1, 0]
2: [1, 0]
object_rooms:
bike: 1
airplane: 2
rubberduck: 3
lego man: 1
toilet: 2
truck: 3
backpack: 1
shoppingcart: 1
fridge: 2
soda can: 3
basket: 2
lamp: 3
red door: [3, 1]
green door: [1, 2]
minimal_plan: [['rotate', 90], ['jumpto', 'green door'], ['rotate', 90], ['jumpto', 'airplane'], ['rotate', 180]]
final_observation: [{'name': 'basket', 'direction': 'front-right', 'distance': 'slightly far', 'orientation': None}, {'name': 'green door', 'direction': 'front-slight-left', 'distance': 'mid distance', 'orientation': 'on left wall'}]
task_type: View2ActionTextEvaluationTask
choices: []
kwargs:
(none)
error: target_not_visible
score: 0
evaluation_mode: prompt_cogmap
Evaluation state
Q2: View2ActionTextEvaluationTask

Question 3 (ID: bf8d3e056f97abbf)

✅ Evaluation Results
id: bf8d3e056f97abbf
question: You return to your starting position and face north. Then you have executed an action sequence and changed to a new location and facing direction. You observe the following: fridge is at front-right, mid distance What action sequence led to this final view? The action sequence must be valid and only contain move actions. Answer format: <sequence of move actions> Example: JumpTo(lamp), Rotate(90)
action: None
answer:
final_pos: [19, 4]
final_ori: [0, -1]
room_id: 2
init_pos: [12, 5]
init_ori: [0, 1]
object_positions:
bike: [9, 6]
airplane: [15, 2]
rubberduck: [1, 6]
lego man: [9, 8]
toilet: [19, 4]
truck: [5, 6]
backpack: [11, 3]
shoppingcart: [13, 3]
fridge: [17, 2]
soda can: [5, 8]
basket: [17, 6]
lamp: [1, 10]
red door: [7, 5]
green door: [14, 5]
object_orientations:
bike: [0, -1]
airplane: [0, -1]
rubberduck: [0, 1]
lego man: [0, -1]
toilet: [1, 0]
truck: [1, 0]
backpack: [0, -1]
shoppingcart: [-1, 0]
red door: [1, 0]
green door: [1, 0]
gate_info:
red door:
room_ids: [3, 1]
ori_by_room:
3: [-1, 0]
1: [1, 0]
green door:
room_ids: [1, 2]
ori_by_room:
1: [-1, 0]
2: [1, 0]
object_rooms:
bike: 1
airplane: 2
rubberduck: 3
lego man: 1
toilet: 2
truck: 3
backpack: 1
shoppingcart: 1
fridge: 2
soda can: 3
basket: 2
lamp: 3
red door: [3, 1]
green door: [1, 2]
minimal_plan: [['rotate', 90], ['jumpto', 'green door'], ['jumpto', 'toilet'], ['rotate', 90]]
final_observation: [{'name': 'fridge', 'direction': 'front-right', 'distance': 'mid distance', 'orientation': None}]
task_type: View2ActionTextEvaluationTask
choices: []
kwargs:
(none)
pos_match: False
ori_match: False
final_pos: [14.0, 5.0]
final_ori: [1, 0]
visible_match: False
score: 0
evaluation_mode: prompt_cogmap
Evaluation state
Q3: View2ActionTextEvaluationTask

📊 Task: AlloMappingEvaluationTask

Question 1 (ID: 11808d88af254359)

✅ Evaluation Results
id: 11808d88af254359
question: Treat your starting position as the origin (0, 0) while facing north. Report allocentric coordinates using (x right/east, y up/north). Objects: basket, toilet, fridge. Answer format: (x0, y0); (x1, y1); ... in the same order. Example: (1, 0); (-2, 3); (0, -1)
action: None
answer: [[5, 1], [7, -1], [5, -3]]
task_type: AlloMappingEvaluationTask
choices: []
kwargs:
(none)
similarity: 0.871
threshold: 0.900
score: 0.871
evaluation_mode: prompt_cogmap
Evaluation state
Q1: AlloMappingEvaluationTask

Question 2 (ID: 2b6ee8e5881229c1)

✅ Evaluation Results
id: 2b6ee8e5881229c1
question: Treat your starting position as the origin (0, 0) while facing north. Report allocentric coordinates using (x right/east, y up/north). Objects: rubberduck, lego man, basket, bike, airplane. Answer format: (x0, y0); (x1, y1); ... in the same order. Example: (1, 0); (-2, 3); (0, -1)
action: None
answer: [[-11, 1], [-3, 3], [5, 1], [-3, 1], [3, -3]]
task_type: AlloMappingEvaluationTask
choices: []
kwargs:
(none)
similarity: 0.893
threshold: 0.900
score: 0.893
evaluation_mode: prompt_cogmap
Evaluation state
Q2: AlloMappingEvaluationTask

Question 3 (ID: 6a481cb096f39335)

✅ Evaluation Results
id: 6a481cb096f39335
question: Treat your starting position as the origin (0, 0) while facing north. Report allocentric coordinates using (x right/east, y up/north). Objects: backpack, basket, lamp, airplane, rubberduck. Answer format: (x0, y0); (x1, y1); ... in the same order. Example: (1, 0); (-2, 3); (0, -1)
action: None
answer: [[-1, -2], [5, 1], [-11, 5], [3, -3], [-11, 1]]
task_type: AlloMappingEvaluationTask
choices: []
kwargs:
(none)
similarity: 0.914
threshold: 0.900
score: 0.914
evaluation_mode: prompt_cogmap
Evaluation state
Q3: AlloMappingEvaluationTask

📊 Task: RotEvaluationTask

Question 1 (ID: 37b10f92ed55a226)

✅ Evaluation Results
id: 37b10f92ed55a226
question: You return to your starting position and face north. You will perform a full 360-degree rotation by continuously turning clockwise in place. Assume all walls are removed (you can see through walls), so every object is visible. Focus on this set of objects: basket, fridge, shoppingcart. List them in the exact order they appear directly ahead while you rotate. If two objects share a bearing, place the nearer one first. Answer format: <object_name1>, <object_name2>, ... Example: mug, sofa, plant
action: None
answer: ['basket', 'fridge', 'shoppingcart']
task_type: RotEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: prompt_cogmap
Evaluation state
Q1: RotEvaluationTask

Question 2 (ID: 4b2dc57e77686a8b)

✅ Evaluation Results
id: 4b2dc57e77686a8b
question: You return to your starting position and face north. You will perform a full 360-degree rotation by continuously turning counterclockwise in place. Assume all walls are removed (you can see through walls), so every object is visible. Focus on this set of objects: backpack, lego man, rubberduck. List them in the exact order they appear directly ahead while you rotate. If two objects share a bearing, place the nearer one first. Answer format: <object_name1>, <object_name2>, ... Example: mug, sofa, plant
action: None
answer: ['lego man', 'rubberduck', 'backpack']
task_type: RotEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: prompt_cogmap
Evaluation state
Q2: RotEvaluationTask

Question 3 (ID: 3af660d5e0016740)

✅ Evaluation Results
id: 3af660d5e0016740
question: You return to your starting position and face north. You will perform a full 360-degree rotation by continuously turning clockwise in place. Assume all walls are removed (you can see through walls), so every object is visible. Focus on this set of objects: airplane, backpack, lamp, toilet. List them in the exact order they appear directly ahead while you rotate. If two objects share a bearing, place the nearer one first. Answer format: <object_name1>, <object_name2>, ... Example: mug, sofa, plant
action: None
answer: ['toilet', 'airplane', 'backpack', 'lamp']
task_type: RotEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: prompt_cogmap
Evaluation state
Q3: RotEvaluationTask

📊 Task: Location2ViewEvaluationTask

Question 1 (ID: 3af268195886a8df)

✅ Evaluation Results
id: 3af268195886a8df
question: Treat the green door as the new 'origin' (0, 0). You move to (2, -2) and face east. What is the egocentric relation of toilet? Answer format: <direction>, <distance> Example: front, near
action: None
answer: front-slight-left, mid distance
task_type: Location2ViewEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: prompt_cogmap
Evaluation state
Q1: Location2ViewEvaluationTask

Question 2 (ID: 56728a7c1f765ce3)

✅ Evaluation Results
id: 56728a7c1f765ce3
question: Treat the green door as the new 'origin' (0, 0). You move to (2, -2) and face north. What is the egocentric relation of basket? Answer format: <direction>, <distance> Example: front, near
action: None
answer: front-slight-right, mid distance
task_type: Location2ViewEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: prompt_cogmap
Evaluation state
Q2: Location2ViewEvaluationTask

Question 3 (ID: 843e8e1b199cd022)

✅ Evaluation Results
id: 843e8e1b199cd022
question: Treat the red door as the new 'origin' (0, 0). You move to (-3, 4) and face south. What is the egocentric relation of truck? Answer format: <direction>, <distance> Example: front, near
action: None
answer: front-slight-left, mid distance
task_type: Location2ViewEvaluationTask
choices: []
kwargs:
(none)
score: 0.500
evaluation_mode: prompt_cogmap
Evaluation state
Q3: Location2ViewEvaluationTask

📊 Task: View2LocationTextEvaluationTask

Question 1 (ID: 0627604bb51f9403)

✅ Evaluation Results
id: 0627604bb51f9403
question: You move to a new location and face east. You observe: toilet is at front-slight-left, mid distance, facing forward; fridge is at front-right, near Treat the green door as the new 'origin' (0, 0). What is your new 2D coordinate (x, y)? Answer format: (x, y) Example: (2, -1)
action: None
answer:
coord: [2, -2]
final_pos: [16, 3]
final_ori: [1, 0]
room_id: 2
object_positions:
bike: [9, 6]
airplane: [15, 2]
rubberduck: [1, 6]
lego man: [9, 8]
toilet: [19, 4]
truck: [5, 6]
backpack: [11, 3]
shoppingcart: [13, 3]
fridge: [17, 2]
soda can: [5, 8]
basket: [17, 6]
lamp: [1, 10]
red door: [7, 5]
green door: [14, 5]
object_orientations:
bike: [0, -1]
airplane: [0, -1]
rubberduck: [0, 1]
lego man: [0, -1]
toilet: [1, 0]
truck: [1, 0]
backpack: [0, -1]
shoppingcart: [-1, 0]
red door: [1, 0]
green door: [1, 0]
final_observation: [{'name': 'toilet', 'direction': 'front-slight-left', 'distance': 'mid distance', 'orientation': 'facing forward'}, {'name': 'fridge', 'direction': 'front-right', 'distance': 'near', 'orientation': None}]
gate_info:
red door:
room_ids: [3, 1]
ori_by_room:
3: [-1, 0]
1: [1, 0]
green door:
room_ids: [1, 2]
ori_by_room:
1: [-1, 0]
2: [1, 0]
object_rooms:
bike: 1
airplane: 2
rubberduck: 3
lego man: 1
toilet: 2
truck: 3
backpack: 1
shoppingcart: 1
fridge: 2
soda can: 3
basket: 2
lamp: 3
red door: [3, 1]
green door: [1, 2]
task_type: View2LocationTextEvaluationTask
choices: []
kwargs:
(none)
similarity: 1
raw_score: 1
score: 1
evaluation_mode: prompt_cogmap
Evaluation state
Q1: View2LocationTextEvaluationTask

Question 2 (ID: 9d51c9ea1c03280c)

✅ Evaluation Results
id: 9d51c9ea1c03280c
question: You move to a new location and face north. You observe: basket is at front-slight-right, mid distance; green door is at front-left, mid distance, on left wall Treat the green door as the new 'origin' (0, 0). What is your new 2D coordinate (x, y)? Answer format: (x, y) Example: (2, -1)
action: None
answer:
coord: [2, -2]
final_pos: [16, 3]
final_ori: [0, 1]
room_id: 2
object_positions:
bike: [9, 6]
airplane: [15, 2]
rubberduck: [1, 6]
lego man: [9, 8]
toilet: [19, 4]
truck: [5, 6]
backpack: [11, 3]
shoppingcart: [13, 3]
fridge: [17, 2]
soda can: [5, 8]
basket: [17, 6]
lamp: [1, 10]
red door: [7, 5]
green door: [14, 5]
object_orientations:
bike: [0, -1]
airplane: [0, -1]
rubberduck: [0, 1]
lego man: [0, -1]
toilet: [1, 0]
truck: [1, 0]
backpack: [0, -1]
shoppingcart: [-1, 0]
red door: [1, 0]
green door: [1, 0]
final_observation: [{'name': 'basket', 'direction': 'front-slight-right', 'distance': 'mid distance', 'orientation': None}, {'name': 'green door', 'direction': 'front-left', 'distance': 'mid distance', 'orientation': 'on left wall'}]
gate_info:
red door:
room_ids: [3, 1]
ori_by_room:
3: [-1, 0]
1: [1, 0]
green door:
room_ids: [1, 2]
ori_by_room:
1: [-1, 0]
2: [1, 0]
object_rooms:
bike: 1
airplane: 2
rubberduck: 3
lego man: 1
toilet: 2
truck: 3
backpack: 1
shoppingcart: 1
fridge: 2
soda can: 3
basket: 2
lamp: 3
red door: [3, 1]
green door: [1, 2]
task_type: View2LocationTextEvaluationTask
choices: []
kwargs:
(none)
similarity: 1
raw_score: 1
score: 1
evaluation_mode: prompt_cogmap
Evaluation state
Q2: View2LocationTextEvaluationTask

Question 3 (ID: 231594f7927a5d40)

✅ Evaluation Results
id: 231594f7927a5d40
question: You move to a new location and face south. You observe: truck is at front-slight-left, mid distance, facing left; rubberduck is at front-right, slightly far, facing backward; red door is at front-left, slightly far, on left wall Treat the red door as the new 'origin' (0, 0). What is your new 2D coordinate (x, y)? Answer format: (x, y) Example: (2, -1)
action: None
answer:
coord: [-3, 4]
final_pos: [4, 9]
final_ori: [0, -1]
room_id: 3
object_positions:
bike: [9, 6]
airplane: [15, 2]
rubberduck: [1, 6]
lego man: [9, 8]
toilet: [19, 4]
truck: [5, 6]
backpack: [11, 3]
shoppingcart: [13, 3]
fridge: [17, 2]
soda can: [5, 8]
basket: [17, 6]
lamp: [1, 10]
red door: [7, 5]
green door: [14, 5]
object_orientations:
bike: [0, -1]
airplane: [0, -1]
rubberduck: [0, 1]
lego man: [0, -1]
toilet: [1, 0]
truck: [1, 0]
backpack: [0, -1]
shoppingcart: [-1, 0]
red door: [1, 0]
green door: [1, 0]
final_observation: [{'name': 'truck', 'direction': 'front-slight-left', 'distance': 'mid distance', 'orientation': 'facing left'}, {'name': 'rubberduck', 'direction': 'front-right', 'distance': 'slightly far', 'orientation': 'facing backward'}, {'name': 'red door', 'direction': 'front-left', 'distance': 'slightly far', 'orientation': 'on left wall'}]
gate_info:
red door:
room_ids: [3, 1]
ori_by_room:
3: [-1, 0]
1: [1, 0]
green door:
room_ids: [1, 2]
ori_by_room:
1: [-1, 0]
2: [1, 0]
object_rooms:
bike: 1
airplane: 2
rubberduck: 3
lego man: 1
toilet: 2
truck: 3
backpack: 1
shoppingcart: 1
fridge: 2
soda can: 3
basket: 2
lamp: 3
red door: [3, 1]
green door: [1, 2]
task_type: View2LocationTextEvaluationTask
choices: []
kwargs:
(none)
similarity: 1
raw_score: 1
score: 1
evaluation_mode: prompt_cogmap
Evaluation state
Q3: View2LocationTextEvaluationTask

📊 Evaluation (use_gt_cogmap)

📊 Task: DirectionEvaluationTask

Question 1 (ID: 9d13d2040c8f0e75)

✅ Evaluation Results
id: 9d13d2040c8f0e75
question: You return to your starting position and face north. From a Top-Down map, describe where basket is relative to toilet. Answer format: <cardinal direction>, <distance> Example: north-west, near
action: None
answer: north west, mid distance
task_type: DirectionEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: use_gt_cogmap
Evaluation state
Q1: DirectionEvaluationTask

Question 2 (ID: e2300d26c5895d2d)

✅ Evaluation Results
id: e2300d26c5895d2d
question: You return to your starting position and face north. From a Top-Down map, describe where rubberduck is relative to lego man. Answer format: <cardinal direction>, <distance> Example: north-west, near
action: None
answer: west, far
task_type: DirectionEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: use_gt_cogmap
Evaluation state
Q2: DirectionEvaluationTask

Question 3 (ID: c2a21d8364e4dca4)

✅ Evaluation Results
id: c2a21d8364e4dca4
question: You return to your starting position and face north. From a Top-Down map, describe where shoppingcart is relative to airplane. Answer format: <cardinal direction>, <distance> Example: north-west, near
action: None
answer: north west, mid distance
task_type: DirectionEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: use_gt_cogmap
Evaluation state
Q3: DirectionEvaluationTask

📊 Task: PovEvaluationTask

Question 1 (ID: 063960e28e84e1b8)

✅ Evaluation Results
id: 063960e28e84e1b8
question: Now you jump to rubberduck's direction, facing its direction. Describe where lamp is relative to you. Answer format: <ego direction>, <distance> Example: front-left, near
action: None
answer: front, mid distance
task_type: PovEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: use_gt_cogmap
Evaluation state
Q1: PovEvaluationTask

Question 2 (ID: dbbb7b93fba9d779)

✅ Evaluation Results
id: dbbb7b93fba9d779
question: Now you jump to lego man's direction, facing its direction. Describe where shoppingcart is relative to you. Answer format: <ego direction>, <distance> Example: front-left, near
action: None
answer: front-left, slightly far
task_type: PovEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: use_gt_cogmap
Evaluation state
Q2: PovEvaluationTask

Question 3 (ID: b2e43f1e4b97a13e)

✅ Evaluation Results
id: b2e43f1e4b97a13e
question: Now you jump to shoppingcart's direction, facing its direction. Describe where bike is relative to you. Answer format: <ego direction>, <distance> Example: front-left, near
action: None
answer: front-right, slightly far
task_type: PovEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: use_gt_cogmap
Evaluation state
Q3: PovEvaluationTask

📊 Task: BackwardPovTextEvaluationTask

Question 1 (ID: 4022a79578213eb8)

✅ Evaluation Results
id: 4022a79578213eb8
question: Now you jump to an object's position, facing its direction. You observe that shoppingcart is front-left, slightly far, facing right; red door is front-right, mid distance, on right wall; backpack is front-slight-left, slightly far, facing forward. Which object are you standing at? Answer format: <object_name> Example: lamp
action: None
answer:
answer: lego man
final_pos: [9, 8]
final_ori: [0, -1]
final_observation: [{'name': 'shoppingcart', 'direction': 'front-left', 'distance': 'slightly far', 'orientation': 'facing right'}, {'name': 'red door', 'direction': 'front-right', 'distance': 'mid distance', 'orientation': 'on right wall'}, {'name': 'backpack', 'direction': 'front-slight-left', 'distance': 'slightly far', 'orientation': 'facing forward'}]
object_positions:
bike: [9.0, 6.0]
airplane: [15.0, 2.0]
rubberduck: [1.0, 6.0]
lego man: [9.0, 8.0]
toilet: [19.0, 4.0]
truck: [5.0, 6.0]
backpack: [11.0, 3.0]
shoppingcart: [13.0, 3.0]
fridge: [17.0, 2.0]
soda can: [5.0, 8.0]
basket: [17.0, 6.0]
lamp: [1.0, 10.0]
red door: [7.0, 5.0]
green door: [14.0, 5.0]
object_orientations:
bike: [0, -1]
airplane: [0, -1]
rubberduck: [0, 1]
lego man: [0, -1]
toilet: [1, 0]
truck: [1, 0]
backpack: [0, -1]
shoppingcart: [-1, 0]
red door: [1, 0]
green door: [1, 0]
room_id: 1
gate_info:
red door:
room_ids: [3, 1]
ori_by_room:
3: [-1, 0]
1: [1, 0]
green door:
room_ids: [1, 2]
ori_by_room:
1: [-1, 0]
2: [1, 0]
object_rooms:
bike: 1
airplane: 2
rubberduck: 3
lego man: 1
toilet: 2
truck: 3
backpack: 1
shoppingcart: 1
fridge: 2
soda can: 3
basket: 2
lamp: 3
red door: [3, 1]
green door: [1, 2]
task_type: BackwardPovTextEvaluationTask
choices: []
kwargs:
(none)
name_match: True
view_match: True
score: 1
evaluation_mode: use_gt_cogmap
Evaluation state
Q1: BackwardPovTextEvaluationTask

Question 2 (ID: d217d8691521707e)

✅ Evaluation Results
id: d217d8691521707e
question: Now you jump to an object's position, facing its direction. You observe that red door is front-right, mid distance, on front wall. Which object are you standing at? Answer format: <object_name> Example: lamp
action: None
answer:
answer: truck
final_pos: [5, 6]
final_ori: [1, 0]
final_observation: [{'name': 'red door', 'direction': 'front-right', 'distance': 'mid distance', 'orientation': 'on front wall'}]
object_positions:
bike: [9.0, 6.0]
airplane: [15.0, 2.0]
rubberduck: [1.0, 6.0]
lego man: [9.0, 8.0]
toilet: [19.0, 4.0]
truck: [5.0, 6.0]
backpack: [11.0, 3.0]
shoppingcart: [13.0, 3.0]
fridge: [17.0, 2.0]
soda can: [5.0, 8.0]
basket: [17.0, 6.0]
lamp: [1.0, 10.0]
red door: [7.0, 5.0]
green door: [14.0, 5.0]
object_orientations:
bike: [0, -1]
airplane: [0, -1]
rubberduck: [0, 1]
lego man: [0, -1]
toilet: [1, 0]
truck: [1, 0]
backpack: [0, -1]
shoppingcart: [-1, 0]
red door: [1, 0]
green door: [1, 0]
room_id: 3
gate_info:
red door:
room_ids: [3, 1]
ori_by_room:
3: [-1, 0]
1: [1, 0]
green door:
room_ids: [1, 2]
ori_by_room:
1: [-1, 0]
2: [1, 0]
object_rooms:
bike: 1
airplane: 2
rubberduck: 3
lego man: 1
toilet: 2
truck: 3
backpack: 1
shoppingcart: 1
fridge: 2
soda can: 3
basket: 2
lamp: 3
red door: [3, 1]
green door: [1, 2]
task_type: BackwardPovTextEvaluationTask
choices: []
kwargs:
(none)
name_match: True
view_match: True
score: 1
evaluation_mode: use_gt_cogmap
Evaluation state
Q2: BackwardPovTextEvaluationTask

Question 3 (ID: 7599c12dc388ac1a)

✅ Evaluation Results
id: 7599c12dc388ac1a
question: Now you jump to an object's position, facing its direction. You observe that lamp is front, mid distance. Which object are you standing at? Answer format: <object_name> Example: lamp
action: None
answer:
answer: rubberduck
final_pos: [1, 6]
final_ori: [0, 1]
final_observation: [{'name': 'lamp', 'direction': 'front', 'distance': 'mid distance', 'orientation': None}]
object_positions:
bike: [9.0, 6.0]
airplane: [15.0, 2.0]
rubberduck: [1.0, 6.0]
lego man: [9.0, 8.0]
toilet: [19.0, 4.0]
truck: [5.0, 6.0]
backpack: [11.0, 3.0]
shoppingcart: [13.0, 3.0]
fridge: [17.0, 2.0]
soda can: [5.0, 8.0]
basket: [17.0, 6.0]
lamp: [1.0, 10.0]
red door: [7.0, 5.0]
green door: [14.0, 5.0]
object_orientations:
bike: [0, -1]
airplane: [0, -1]
rubberduck: [0, 1]
lego man: [0, -1]
toilet: [1, 0]
truck: [1, 0]
backpack: [0, -1]
shoppingcart: [-1, 0]
red door: [1, 0]
green door: [1, 0]
room_id: 3
gate_info:
red door:
room_ids: [3, 1]
ori_by_room:
3: [-1, 0]
1: [1, 0]
green door:
room_ids: [1, 2]
ori_by_room:
1: [-1, 0]
2: [1, 0]
object_rooms:
bike: 1
airplane: 2
rubberduck: 3
lego man: 1
toilet: 2
truck: 3
backpack: 1
shoppingcart: 1
fridge: 2
soda can: 3
basket: 2
lamp: 3
red door: [3, 1]
green door: [1, 2]
task_type: BackwardPovTextEvaluationTask
choices: []
kwargs:
(none)
name_match: True
view_match: True
score: 1
evaluation_mode: use_gt_cogmap
Evaluation state
Q3: BackwardPovTextEvaluationTask

📊 Task: Action2ViewEvaluationTask

Question 1 (ID: e5146a0f564c765f)

✅ Evaluation Results
id: e5146a0f564c765f
question: You return to your starting position and face north. You will execute the following action sequence: 1. Rotate(90) 2. Jump to the object at front, near. 3. Jump to the object at front-right, slightly far. After executing the actions, what is the ego relation of toilet relative to you? Answer format: <ego direction>, <distance> Example: front, near
action: None
answer: front-left, mid distance
task_type: Action2ViewEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: use_gt_cogmap
Evaluation state
Q1: Action2ViewEvaluationTask

Question 2 (ID: 90f3cb1a03f64a70)

✅ Evaluation Results
id: 90f3cb1a03f64a70
question: You return to your starting position and face north. You will execute the following action sequence: 1. Rotate(90) 2. Jump to the object at front, near. 3. Jump to the object at front-right, slightly far. 4. Rotate(-180) After executing the actions, what is the ego relation of green door relative to you? Answer format: <ego direction>, <distance> Example: front, near
action: None
answer: front-right, slightly far
task_type: Action2ViewEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: use_gt_cogmap
Evaluation state
Q2: Action2ViewEvaluationTask

Question 3 (ID: 7b132df57af38164)

✅ Evaluation Results
id: 7b132df57af38164
question: You return to your starting position and face north. You will execute the following action sequence: 1. Jump to the object at front-left, slightly far. 2. Rotate(90) 3. Jump to the object at front-right, slightly far. 4. Rotate(90) After executing the actions, what is the ego relation of fridge relative to you? Answer format: <ego direction>, <distance> Example: front, near
action: None
answer: front-left, slightly far
task_type: Action2ViewEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: use_gt_cogmap
Evaluation state
Q3: Action2ViewEvaluationTask

📊 Task: View2ActionTextEvaluationTask

Question 1 (ID: a1b55f183063b5a2)

✅ Evaluation Results
id: a1b55f183063b5a2
question: You return to your starting position and face north. Then you have executed an action sequence and changed to a new location and facing direction. You observe the following: truck is at front, near, facing left; red door is at front-left, mid distance, on left wall What action sequence led to this final view? The action sequence must be valid and only contain move actions. Answer format: <sequence of move actions> Example: JumpTo(lamp), Rotate(90)
action: None
answer:
final_pos: [5, 8]
final_ori: [0, -1]
room_id: 3
init_pos: [12, 5]
init_ori: [0, 1]
object_positions:
bike: [9, 6]
airplane: [15, 2]
rubberduck: [1, 6]
lego man: [9, 8]
toilet: [19, 4]
truck: [5, 6]
backpack: [11, 3]
shoppingcart: [13, 3]
fridge: [17, 2]
soda can: [5, 8]
basket: [17, 6]
lamp: [1, 10]
red door: [7, 5]
green door: [14, 5]
object_orientations:
bike: [0, -1]
airplane: [0, -1]
rubberduck: [0, 1]
lego man: [0, -1]
toilet: [1, 0]
truck: [1, 0]
backpack: [0, -1]
shoppingcart: [-1, 0]
red door: [1, 0]
green door: [1, 0]
gate_info:
red door:
room_ids: [3, 1]
ori_by_room:
3: [-1, 0]
1: [1, 0]
green door:
room_ids: [1, 2]
ori_by_room:
1: [-1, 0]
2: [1, 0]
object_rooms:
bike: 1
airplane: 2
rubberduck: 3
lego man: 1
toilet: 2
truck: 3
backpack: 1
shoppingcart: 1
fridge: 2
soda can: 3
basket: 2
lamp: 3
red door: [3, 1]
green door: [1, 2]
minimal_plan: [['rotate', -90], ['jumpto', 'red door'], ['rotate', 90], ['jumpto', 'soda can'], ['rotate', 180]]
final_observation: [{'name': 'truck', 'direction': 'front', 'distance': 'near', 'orientation': 'facing left'}, {'name': 'red door', 'direction': 'front-left', 'distance': 'mid distance', 'orientation': 'on left wall'}]
task_type: View2ActionTextEvaluationTask
choices: []
kwargs:
(none)
pos_match: True
ori_match: True
final_pos: [5.0, 8.0]
final_ori: [0, -1]
visible_match: True
score: 1
evaluation_mode: use_gt_cogmap
Evaluation state
Q1: View2ActionTextEvaluationTask

Question 2 (ID: 514e1fc99b9dfec8)

✅ Evaluation Results
id: 514e1fc99b9dfec8
question: You return to your starting position and face north. Then you have executed an action sequence and changed to a new location and facing direction. You observe the following: basket is at front-right, slightly far; green door is at front-slight-left, mid distance, on left wall What action sequence led to this final view? The action sequence must be valid and only contain move actions. Answer format: <sequence of move actions> Example: JumpTo(lamp), Rotate(90)
action: None
answer:
final_pos: [15, 2]
final_ori: [0, 1]
room_id: 2
init_pos: [12, 5]
init_ori: [0, 1]
object_positions:
bike: [9, 6]
airplane: [15, 2]
rubberduck: [1, 6]
lego man: [9, 8]
toilet: [19, 4]
truck: [5, 6]
backpack: [11, 3]
shoppingcart: [13, 3]
fridge: [17, 2]
soda can: [5, 8]
basket: [17, 6]
lamp: [1, 10]
red door: [7, 5]
green door: [14, 5]
object_orientations:
bike: [0, -1]
airplane: [0, -1]
rubberduck: [0, 1]
lego man: [0, -1]
toilet: [1, 0]
truck: [1, 0]
backpack: [0, -1]
shoppingcart: [-1, 0]
red door: [1, 0]
green door: [1, 0]
gate_info:
red door:
room_ids: [3, 1]
ori_by_room:
3: [-1, 0]
1: [1, 0]
green door:
room_ids: [1, 2]
ori_by_room:
1: [-1, 0]
2: [1, 0]
object_rooms:
bike: 1
airplane: 2
rubberduck: 3
lego man: 1
toilet: 2
truck: 3
backpack: 1
shoppingcart: 1
fridge: 2
soda can: 3
basket: 2
lamp: 3
red door: [3, 1]
green door: [1, 2]
minimal_plan: [['rotate', 90], ['jumpto', 'green door'], ['rotate', 90], ['jumpto', 'airplane'], ['rotate', 180]]
final_observation: [{'name': 'basket', 'direction': 'front-right', 'distance': 'slightly far', 'orientation': None}, {'name': 'green door', 'direction': 'front-slight-left', 'distance': 'mid distance', 'orientation': 'on left wall'}]
task_type: View2ActionTextEvaluationTask
choices: []
kwargs:
(none)
pos_match: True
ori_match: True
final_pos: [15.0, 2.0]
final_ori: [0, 1]
visible_match: True
score: 1
evaluation_mode: use_gt_cogmap
Evaluation state
Q2: View2ActionTextEvaluationTask

Question 3 (ID: bf8d3e056f97abbf)

✅ Evaluation Results
id: bf8d3e056f97abbf
question: You return to your starting position and face north. Then you have executed an action sequence and changed to a new location and facing direction. You observe the following: fridge is at front-right, mid distance What action sequence led to this final view? The action sequence must be valid and only contain move actions. Answer format: <sequence of move actions> Example: JumpTo(lamp), Rotate(90)
action: None
answer:
final_pos: [19, 4]
final_ori: [0, -1]
room_id: 2
init_pos: [12, 5]
init_ori: [0, 1]
object_positions:
bike: [9, 6]
airplane: [15, 2]
rubberduck: [1, 6]
lego man: [9, 8]
toilet: [19, 4]
truck: [5, 6]
backpack: [11, 3]
shoppingcart: [13, 3]
fridge: [17, 2]
soda can: [5, 8]
basket: [17, 6]
lamp: [1, 10]
red door: [7, 5]
green door: [14, 5]
object_orientations:
bike: [0, -1]
airplane: [0, -1]
rubberduck: [0, 1]
lego man: [0, -1]
toilet: [1, 0]
truck: [1, 0]
backpack: [0, -1]
shoppingcart: [-1, 0]
red door: [1, 0]
green door: [1, 0]
gate_info:
red door:
room_ids: [3, 1]
ori_by_room:
3: [-1, 0]
1: [1, 0]
green door:
room_ids: [1, 2]
ori_by_room:
1: [-1, 0]
2: [1, 0]
object_rooms:
bike: 1
airplane: 2
rubberduck: 3
lego man: 1
toilet: 2
truck: 3
backpack: 1
shoppingcart: 1
fridge: 2
soda can: 3
basket: 2
lamp: 3
red door: [3, 1]
green door: [1, 2]
minimal_plan: [['rotate', 90], ['jumpto', 'green door'], ['jumpto', 'toilet'], ['rotate', 90]]
final_observation: [{'name': 'fridge', 'direction': 'front-right', 'distance': 'mid distance', 'orientation': None}]
task_type: View2ActionTextEvaluationTask
choices: []
kwargs:
(none)
pos_match: True
ori_match: True
final_pos: [19.0, 4.0]
final_ori: [0, -1]
visible_match: True
score: 1
evaluation_mode: use_gt_cogmap
Evaluation state
Q3: View2ActionTextEvaluationTask

📊 Task: AlloMappingEvaluationTask

Question 1 (ID: 11808d88af254359)

✅ Evaluation Results
id: 11808d88af254359
question: Treat your starting position as the origin (0, 0) while facing north. Report allocentric coordinates using (x right/east, y up/north). Objects: basket, toilet, fridge. Answer format: (x0, y0); (x1, y1); ... in the same order. Example: (1, 0); (-2, 3); (0, -1)
action: None
answer: [[5, 1], [7, -1], [5, -3]]
task_type: AlloMappingEvaluationTask
choices: []
kwargs:
(none)
similarity: 1
threshold: 0.900
score: 1
evaluation_mode: use_gt_cogmap
Evaluation state
Q1: AlloMappingEvaluationTask

Question 2 (ID: 2b6ee8e5881229c1)

✅ Evaluation Results
id: 2b6ee8e5881229c1
question: Treat your starting position as the origin (0, 0) while facing north. Report allocentric coordinates using (x right/east, y up/north). Objects: rubberduck, lego man, basket, bike, airplane. Answer format: (x0, y0); (x1, y1); ... in the same order. Example: (1, 0); (-2, 3); (0, -1)
action: None
answer: [[-11, 1], [-3, 3], [5, 1], [-3, 1], [3, -3]]
task_type: AlloMappingEvaluationTask
choices: []
kwargs:
(none)
similarity: 1
threshold: 0.900
score: 1
evaluation_mode: use_gt_cogmap
Evaluation state
Q2: AlloMappingEvaluationTask

Question 3 (ID: 6a481cb096f39335)

✅ Evaluation Results
id: 6a481cb096f39335
question: Treat your starting position as the origin (0, 0) while facing north. Report allocentric coordinates using (x right/east, y up/north). Objects: backpack, basket, lamp, airplane, rubberduck. Answer format: (x0, y0); (x1, y1); ... in the same order. Example: (1, 0); (-2, 3); (0, -1)
action: None
answer: [[-1, -2], [5, 1], [-11, 5], [3, -3], [-11, 1]]
task_type: AlloMappingEvaluationTask
choices: []
kwargs:
(none)
similarity: 1
threshold: 0.900
score: 1
evaluation_mode: use_gt_cogmap
Evaluation state
Q3: AlloMappingEvaluationTask

📊 Task: RotEvaluationTask

Question 1 (ID: 37b10f92ed55a226)

✅ Evaluation Results
id: 37b10f92ed55a226
question: You return to your starting position and face north. You will perform a full 360-degree rotation by continuously turning clockwise in place. Assume all walls are removed (you can see through walls), so every object is visible. Focus on this set of objects: basket, fridge, shoppingcart. List them in the exact order they appear directly ahead while you rotate. If two objects share a bearing, place the nearer one first. Answer format: <object_name1>, <object_name2>, ... Example: mug, sofa, plant
action: None
answer: ['basket', 'fridge', 'shoppingcart']
task_type: RotEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: use_gt_cogmap
Evaluation state
Q1: RotEvaluationTask

Question 2 (ID: 4b2dc57e77686a8b)

✅ Evaluation Results
id: 4b2dc57e77686a8b
question: You return to your starting position and face north. You will perform a full 360-degree rotation by continuously turning counterclockwise in place. Assume all walls are removed (you can see through walls), so every object is visible. Focus on this set of objects: backpack, lego man, rubberduck. List them in the exact order they appear directly ahead while you rotate. If two objects share a bearing, place the nearer one first. Answer format: <object_name1>, <object_name2>, ... Example: mug, sofa, plant
action: None
answer: ['lego man', 'rubberduck', 'backpack']
task_type: RotEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: use_gt_cogmap
Evaluation state
Q2: RotEvaluationTask

Question 3 (ID: 3af660d5e0016740)

✅ Evaluation Results
id: 3af660d5e0016740
question: You return to your starting position and face north. You will perform a full 360-degree rotation by continuously turning clockwise in place. Assume all walls are removed (you can see through walls), so every object is visible. Focus on this set of objects: airplane, backpack, lamp, toilet. List them in the exact order they appear directly ahead while you rotate. If two objects share a bearing, place the nearer one first. Answer format: <object_name1>, <object_name2>, ... Example: mug, sofa, plant
action: None
answer: ['toilet', 'airplane', 'backpack', 'lamp']
task_type: RotEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: use_gt_cogmap
Evaluation state
Q3: RotEvaluationTask

📊 Task: Location2ViewEvaluationTask

Question 1 (ID: 3af268195886a8df)

✅ Evaluation Results
id: 3af268195886a8df
question: Treat the green door as the new 'origin' (0, 0). You move to (2, -2) and face east. What is the egocentric relation of toilet? Answer format: <direction>, <distance> Example: front, near
action: None
answer: front-slight-left, mid distance
task_type: Location2ViewEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: use_gt_cogmap
Evaluation state
Q1: Location2ViewEvaluationTask

Question 2 (ID: 56728a7c1f765ce3)

✅ Evaluation Results
id: 56728a7c1f765ce3
question: Treat the green door as the new 'origin' (0, 0). You move to (2, -2) and face north. What is the egocentric relation of basket? Answer format: <direction>, <distance> Example: front, near
action: None
answer: front-slight-right, mid distance
task_type: Location2ViewEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: use_gt_cogmap
Evaluation state
Q2: Location2ViewEvaluationTask

Question 3 (ID: 843e8e1b199cd022)

✅ Evaluation Results
id: 843e8e1b199cd022
question: Treat the red door as the new 'origin' (0, 0). You move to (-3, 4) and face south. What is the egocentric relation of truck? Answer format: <direction>, <distance> Example: front, near
action: None
answer: front-slight-left, mid distance
task_type: Location2ViewEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: use_gt_cogmap
Evaluation state
Q3: Location2ViewEvaluationTask

📊 Task: View2LocationTextEvaluationTask

Question 1 (ID: 0627604bb51f9403)

✅ Evaluation Results
id: 0627604bb51f9403
question: You move to a new location and face east. You observe: toilet is at front-slight-left, mid distance, facing forward; fridge is at front-right, near Treat the green door as the new 'origin' (0, 0). What is your new 2D coordinate (x, y)? Answer format: (x, y) Example: (2, -1)
action: None
answer:
coord: [2, -2]
final_pos: [16, 3]
final_ori: [1, 0]
room_id: 2
object_positions:
bike: [9, 6]
airplane: [15, 2]
rubberduck: [1, 6]
lego man: [9, 8]
toilet: [19, 4]
truck: [5, 6]
backpack: [11, 3]
shoppingcart: [13, 3]
fridge: [17, 2]
soda can: [5, 8]
basket: [17, 6]
lamp: [1, 10]
red door: [7, 5]
green door: [14, 5]
object_orientations:
bike: [0, -1]
airplane: [0, -1]
rubberduck: [0, 1]
lego man: [0, -1]
toilet: [1, 0]
truck: [1, 0]
backpack: [0, -1]
shoppingcart: [-1, 0]
red door: [1, 0]
green door: [1, 0]
final_observation: [{'name': 'toilet', 'direction': 'front-slight-left', 'distance': 'mid distance', 'orientation': 'facing forward'}, {'name': 'fridge', 'direction': 'front-right', 'distance': 'near', 'orientation': None}]
gate_info:
red door:
room_ids: [3, 1]
ori_by_room:
3: [-1, 0]
1: [1, 0]
green door:
room_ids: [1, 2]
ori_by_room:
1: [-1, 0]
2: [1, 0]
object_rooms:
bike: 1
airplane: 2
rubberduck: 3
lego man: 1
toilet: 2
truck: 3
backpack: 1
shoppingcart: 1
fridge: 2
soda can: 3
basket: 2
lamp: 3
red door: [3, 1]
green door: [1, 2]
task_type: View2LocationTextEvaluationTask
choices: []
kwargs:
(none)
similarity: 1
raw_score: 1
score: 1
evaluation_mode: use_gt_cogmap
Evaluation state
Q1: View2LocationTextEvaluationTask

Question 2 (ID: 9d51c9ea1c03280c)

✅ Evaluation Results
id: 9d51c9ea1c03280c
question: You move to a new location and face north. You observe: basket is at front-slight-right, mid distance; green door is at front-left, mid distance, on left wall Treat the green door as the new 'origin' (0, 0). What is your new 2D coordinate (x, y)? Answer format: (x, y) Example: (2, -1)
action: None
answer:
coord: [2, -2]
final_pos: [16, 3]
final_ori: [0, 1]
room_id: 2
object_positions:
bike: [9, 6]
airplane: [15, 2]
rubberduck: [1, 6]
lego man: [9, 8]
toilet: [19, 4]
truck: [5, 6]
backpack: [11, 3]
shoppingcart: [13, 3]
fridge: [17, 2]
soda can: [5, 8]
basket: [17, 6]
lamp: [1, 10]
red door: [7, 5]
green door: [14, 5]
object_orientations:
bike: [0, -1]
airplane: [0, -1]
rubberduck: [0, 1]
lego man: [0, -1]
toilet: [1, 0]
truck: [1, 0]
backpack: [0, -1]
shoppingcart: [-1, 0]
red door: [1, 0]
green door: [1, 0]
final_observation: [{'name': 'basket', 'direction': 'front-slight-right', 'distance': 'mid distance', 'orientation': None}, {'name': 'green door', 'direction': 'front-left', 'distance': 'mid distance', 'orientation': 'on left wall'}]
gate_info:
red door:
room_ids: [3, 1]
ori_by_room:
3: [-1, 0]
1: [1, 0]
green door:
room_ids: [1, 2]
ori_by_room:
1: [-1, 0]
2: [1, 0]
object_rooms:
bike: 1
airplane: 2
rubberduck: 3
lego man: 1
toilet: 2
truck: 3
backpack: 1
shoppingcart: 1
fridge: 2
soda can: 3
basket: 2
lamp: 3
red door: [3, 1]
green door: [1, 2]
task_type: View2LocationTextEvaluationTask
choices: []
kwargs:
(none)
similarity: 1
raw_score: 1
score: 1
evaluation_mode: use_gt_cogmap
Evaluation state
Q2: View2LocationTextEvaluationTask

Question 3 (ID: 231594f7927a5d40)

✅ Evaluation Results
id: 231594f7927a5d40
question: You move to a new location and face south. You observe: truck is at front-slight-left, mid distance, facing left; rubberduck is at front-right, slightly far, facing backward; red door is at front-left, slightly far, on left wall Treat the red door as the new 'origin' (0, 0). What is your new 2D coordinate (x, y)? Answer format: (x, y) Example: (2, -1)
action: None
answer:
coord: [-3, 4]
final_pos: [4, 9]
final_ori: [0, -1]
room_id: 3
object_positions:
bike: [9, 6]
airplane: [15, 2]
rubberduck: [1, 6]
lego man: [9, 8]
toilet: [19, 4]
truck: [5, 6]
backpack: [11, 3]
shoppingcart: [13, 3]
fridge: [17, 2]
soda can: [5, 8]
basket: [17, 6]
lamp: [1, 10]
red door: [7, 5]
green door: [14, 5]
object_orientations:
bike: [0, -1]
airplane: [0, -1]
rubberduck: [0, 1]
lego man: [0, -1]
toilet: [1, 0]
truck: [1, 0]
backpack: [0, -1]
shoppingcart: [-1, 0]
red door: [1, 0]
green door: [1, 0]
final_observation: [{'name': 'truck', 'direction': 'front-slight-left', 'distance': 'mid distance', 'orientation': 'facing left'}, {'name': 'rubberduck', 'direction': 'front-right', 'distance': 'slightly far', 'orientation': 'facing backward'}, {'name': 'red door', 'direction': 'front-left', 'distance': 'slightly far', 'orientation': 'on left wall'}]
gate_info:
red door:
room_ids: [3, 1]
ori_by_room:
3: [-1, 0]
1: [1, 0]
green door:
room_ids: [1, 2]
ori_by_room:
1: [-1, 0]
2: [1, 0]
object_rooms:
bike: 1
airplane: 2
rubberduck: 3
lego man: 1
toilet: 2
truck: 3
backpack: 1
shoppingcart: 1
fridge: 2
soda can: 3
basket: 2
lamp: 3
red door: [3, 1]
green door: [1, 2]
task_type: View2LocationTextEvaluationTask
choices: []
kwargs:
(none)
similarity: 1
raw_score: 1
score: 1
evaluation_mode: use_gt_cogmap
Evaluation state
Q3: View2LocationTextEvaluationTask

📊 Evaluation (use_model_cogmap)

📊 Task: View2ActionTextEvaluationTask

Question 1 (ID: bf8d3e056f97abbf)

✅ Evaluation Results
id: bf8d3e056f97abbf
question: You return to your starting position and face north. Then you have executed an action sequence and changed to a new location and facing direction. You observe the following: fridge is at front-right, mid distance What action sequence led to this final view? The action sequence must be valid and only contain move actions. Answer format: <sequence of move actions> Example: JumpTo(lamp), Rotate(90)
action: None
answer:
final_pos: [19, 4]
final_ori: [0, -1]
room_id: 2
init_pos: [12, 5]
init_ori: [0, 1]
object_positions:
bike: [9, 6]
airplane: [15, 2]
rubberduck: [1, 6]
lego man: [9, 8]
toilet: [19, 4]
truck: [5, 6]
backpack: [11, 3]
shoppingcart: [13, 3]
fridge: [17, 2]
soda can: [5, 8]
basket: [17, 6]
lamp: [1, 10]
red door: [7, 5]
green door: [14, 5]
object_orientations:
bike: [0, -1]
airplane: [0, -1]
rubberduck: [0, 1]
lego man: [0, -1]
toilet: [1, 0]
truck: [1, 0]
backpack: [0, -1]
shoppingcart: [-1, 0]
red door: [1, 0]
green door: [1, 0]
gate_info:
red door:
room_ids: [3, 1]
ori_by_room:
3: [-1, 0]
1: [1, 0]
green door:
room_ids: [1, 2]
ori_by_room:
1: [-1, 0]
2: [1, 0]
object_rooms:
bike: 1
airplane: 2
rubberduck: 3
lego man: 1
toilet: 2
truck: 3
backpack: 1
shoppingcart: 1
fridge: 2
soda can: 3
basket: 2
lamp: 3
red door: [3, 1]
green door: [1, 2]
minimal_plan: [['rotate', 90], ['jumpto', 'green door'], ['jumpto', 'toilet'], ['rotate', 90]]
final_observation: [{'name': 'fridge', 'direction': 'front-right', 'distance': 'mid distance', 'orientation': None}]
task_type: View2ActionTextEvaluationTask
choices: []
kwargs:
(none)
error: invalid_format
score: 0
evaluation_mode: use_model_cogmap
Evaluation state
Q1: View2ActionTextEvaluationTask

Question 2 (ID: a1b55f183063b5a2)

✅ Evaluation Results
id: a1b55f183063b5a2
question: You return to your starting position and face north. Then you have executed an action sequence and changed to a new location and facing direction. You observe the following: truck is at front, near, facing left; red door is at front-left, mid distance, on left wall What action sequence led to this final view? The action sequence must be valid and only contain move actions. Answer format: <sequence of move actions> Example: JumpTo(lamp), Rotate(90)
action: None
answer:
final_pos: [5, 8]
final_ori: [0, -1]
room_id: 3
init_pos: [12, 5]
init_ori: [0, 1]
object_positions:
bike: [9, 6]
airplane: [15, 2]
rubberduck: [1, 6]
lego man: [9, 8]
toilet: [19, 4]
truck: [5, 6]
backpack: [11, 3]
shoppingcart: [13, 3]
fridge: [17, 2]
soda can: [5, 8]
basket: [17, 6]
lamp: [1, 10]
red door: [7, 5]
green door: [14, 5]
object_orientations:
bike: [0, -1]
airplane: [0, -1]
rubberduck: [0, 1]
lego man: [0, -1]
toilet: [1, 0]
truck: [1, 0]
backpack: [0, -1]
shoppingcart: [-1, 0]
red door: [1, 0]
green door: [1, 0]
gate_info:
red door:
room_ids: [3, 1]
ori_by_room:
3: [-1, 0]
1: [1, 0]
green door:
room_ids: [1, 2]
ori_by_room:
1: [-1, 0]
2: [1, 0]
object_rooms:
bike: 1
airplane: 2
rubberduck: 3
lego man: 1
toilet: 2
truck: 3
backpack: 1
shoppingcart: 1
fridge: 2
soda can: 3
basket: 2
lamp: 3
red door: [3, 1]
green door: [1, 2]
minimal_plan: [['rotate', -90], ['jumpto', 'red door'], ['rotate', 90], ['jumpto', 'soda can'], ['rotate', 180]]
final_observation: [{'name': 'truck', 'direction': 'front', 'distance': 'near', 'orientation': 'facing left'}, {'name': 'red door', 'direction': 'front-left', 'distance': 'mid distance', 'orientation': 'on left wall'}]
task_type: View2ActionTextEvaluationTask
choices: []
kwargs:
(none)
error: target_not_visible
score: 0
evaluation_mode: use_model_cogmap
Evaluation state
Q2: View2ActionTextEvaluationTask

Question 3 (ID: 514e1fc99b9dfec8)

✅ Evaluation Results
id: 514e1fc99b9dfec8
question: You return to your starting position and face north. Then you have executed an action sequence and changed to a new location and facing direction. You observe the following: basket is at front-right, slightly far; green door is at front-slight-left, mid distance, on left wall What action sequence led to this final view? The action sequence must be valid and only contain move actions. Answer format: <sequence of move actions> Example: JumpTo(lamp), Rotate(90)
action: None
answer:
final_pos: [15, 2]
final_ori: [0, 1]
room_id: 2
init_pos: [12, 5]
init_ori: [0, 1]
object_positions:
bike: [9, 6]
airplane: [15, 2]
rubberduck: [1, 6]
lego man: [9, 8]
toilet: [19, 4]
truck: [5, 6]
backpack: [11, 3]
shoppingcart: [13, 3]
fridge: [17, 2]
soda can: [5, 8]
basket: [17, 6]
lamp: [1, 10]
red door: [7, 5]
green door: [14, 5]
object_orientations:
bike: [0, -1]
airplane: [0, -1]
rubberduck: [0, 1]
lego man: [0, -1]
toilet: [1, 0]
truck: [1, 0]
backpack: [0, -1]
shoppingcart: [-1, 0]
red door: [1, 0]
green door: [1, 0]
gate_info:
red door:
room_ids: [3, 1]
ori_by_room:
3: [-1, 0]
1: [1, 0]
green door:
room_ids: [1, 2]
ori_by_room:
1: [-1, 0]
2: [1, 0]
object_rooms:
bike: 1
airplane: 2
rubberduck: 3
lego man: 1
toilet: 2
truck: 3
backpack: 1
shoppingcart: 1
fridge: 2
soda can: 3
basket: 2
lamp: 3
red door: [3, 1]
green door: [1, 2]
minimal_plan: [['rotate', 90], ['jumpto', 'green door'], ['rotate', 90], ['jumpto', 'airplane'], ['rotate', 180]]
final_observation: [{'name': 'basket', 'direction': 'front-right', 'distance': 'slightly far', 'orientation': None}, {'name': 'green door', 'direction': 'front-slight-left', 'distance': 'mid distance', 'orientation': 'on left wall'}]
task_type: View2ActionTextEvaluationTask
choices: []
kwargs:
(none)
pos_match: True
ori_match: True
final_pos: [15.0, 2.0]
final_ori: [0, 1]
visible_match: True
score: 1
evaluation_mode: use_model_cogmap
Evaluation state
Q3: View2ActionTextEvaluationTask

📊 Task: AlloMappingEvaluationTask

Question 1 (ID: 11808d88af254359)

✅ Evaluation Results
id: 11808d88af254359
question: Treat your starting position as the origin (0, 0) while facing north. Report allocentric coordinates using (x right/east, y up/north). Objects: basket, toilet, fridge. Answer format: (x0, y0); (x1, y1); ... in the same order. Example: (1, 0); (-2, 3); (0, -1)
action: None
answer: [[5, 1], [7, -1], [5, -3]]
task_type: AlloMappingEvaluationTask
choices: []
kwargs:
(none)
similarity: 0.854
threshold: 0.900
score: 0.854
evaluation_mode: use_model_cogmap
Evaluation state
Q1: AlloMappingEvaluationTask

Question 2 (ID: 2b6ee8e5881229c1)

✅ Evaluation Results
id: 2b6ee8e5881229c1
question: Treat your starting position as the origin (0, 0) while facing north. Report allocentric coordinates using (x right/east, y up/north). Objects: rubberduck, lego man, basket, bike, airplane. Answer format: (x0, y0); (x1, y1); ... in the same order. Example: (1, 0); (-2, 3); (0, -1)
action: None
answer: [[-11, 1], [-3, 3], [5, 1], [-3, 1], [3, -3]]
task_type: AlloMappingEvaluationTask
choices: []
kwargs:
(none)
similarity: 0.859
threshold: 0.900
score: 0.859
evaluation_mode: use_model_cogmap
Evaluation state
Q2: AlloMappingEvaluationTask

Question 3 (ID: 6a481cb096f39335)

✅ Evaluation Results
id: 6a481cb096f39335
question: Treat your starting position as the origin (0, 0) while facing north. Report allocentric coordinates using (x right/east, y up/north). Objects: backpack, basket, lamp, airplane, rubberduck. Answer format: (x0, y0); (x1, y1); ... in the same order. Example: (1, 0); (-2, 3); (0, -1)
action: None
answer: [[-1, -2], [5, 1], [-11, 5], [3, -3], [-11, 1]]
task_type: AlloMappingEvaluationTask
choices: []
kwargs:
(none)
similarity: 0.852
threshold: 0.900
score: 0.852
evaluation_mode: use_model_cogmap
Evaluation state
Q3: AlloMappingEvaluationTask

📊 Task: RotEvaluationTask

Question 1 (ID: 37b10f92ed55a226)

✅ Evaluation Results
id: 37b10f92ed55a226
question: You return to your starting position and face north. You will perform a full 360-degree rotation by continuously turning clockwise in place. Assume all walls are removed (you can see through walls), so every object is visible. Focus on this set of objects: basket, fridge, shoppingcart. List them in the exact order they appear directly ahead while you rotate. If two objects share a bearing, place the nearer one first. Answer format: <object_name1>, <object_name2>, ... Example: mug, sofa, plant
action: None
answer: ['basket', 'fridge', 'shoppingcart']
task_type: RotEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: use_model_cogmap
Evaluation state
Q1: RotEvaluationTask

Question 2 (ID: 4b2dc57e77686a8b)

✅ Evaluation Results
id: 4b2dc57e77686a8b
question: You return to your starting position and face north. You will perform a full 360-degree rotation by continuously turning counterclockwise in place. Assume all walls are removed (you can see through walls), so every object is visible. Focus on this set of objects: backpack, lego man, rubberduck. List them in the exact order they appear directly ahead while you rotate. If two objects share a bearing, place the nearer one first. Answer format: <object_name1>, <object_name2>, ... Example: mug, sofa, plant
action: None
answer: ['lego man', 'rubberduck', 'backpack']
task_type: RotEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: use_model_cogmap
Evaluation state
Q2: RotEvaluationTask

Question 3 (ID: 3af660d5e0016740)

✅ Evaluation Results
id: 3af660d5e0016740
question: You return to your starting position and face north. You will perform a full 360-degree rotation by continuously turning clockwise in place. Assume all walls are removed (you can see through walls), so every object is visible. Focus on this set of objects: airplane, backpack, lamp, toilet. List them in the exact order they appear directly ahead while you rotate. If two objects share a bearing, place the nearer one first. Answer format: <object_name1>, <object_name2>, ... Example: mug, sofa, plant
action: None
answer: ['toilet', 'airplane', 'backpack', 'lamp']
task_type: RotEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: use_model_cogmap
Evaluation state
Q3: RotEvaluationTask

📊 Task: Location2ViewEvaluationTask

Question 1 (ID: 3af268195886a8df)

✅ Evaluation Results
id: 3af268195886a8df
question: Treat the green door as the new 'origin' (0, 0). You move to (2, -2) and face east. What is the egocentric relation of toilet? Answer format: <direction>, <distance> Example: front, near
action: None
answer: front-slight-left, mid distance
task_type: Location2ViewEvaluationTask
choices: []
kwargs:
(none)
score: 0.500
evaluation_mode: use_model_cogmap
Evaluation state
Q1: Location2ViewEvaluationTask

Question 2 (ID: 56728a7c1f765ce3)

✅ Evaluation Results
id: 56728a7c1f765ce3
question: Treat the green door as the new 'origin' (0, 0). You move to (2, -2) and face north. What is the egocentric relation of basket? Answer format: <direction>, <distance> Example: front, near
action: None
answer: front-slight-right, mid distance
task_type: Location2ViewEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: use_model_cogmap
Evaluation state
Q2: Location2ViewEvaluationTask

Question 3 (ID: 843e8e1b199cd022)

✅ Evaluation Results
id: 843e8e1b199cd022
question: Treat the red door as the new 'origin' (0, 0). You move to (-3, 4) and face south. What is the egocentric relation of truck? Answer format: <direction>, <distance> Example: front, near
action: None
answer: front-slight-left, mid distance
task_type: Location2ViewEvaluationTask
choices: []
kwargs:
(none)
score: 0
evaluation_mode: use_model_cogmap
Evaluation state
Q3: Location2ViewEvaluationTask

📊 Task: View2LocationTextEvaluationTask

Question 1 (ID: 0627604bb51f9403)

✅ Evaluation Results
id: 0627604bb51f9403
question: You move to a new location and face east. You observe: toilet is at front-slight-left, mid distance, facing forward; fridge is at front-right, near Treat the green door as the new 'origin' (0, 0). What is your new 2D coordinate (x, y)? Answer format: (x, y) Example: (2, -1)
action: None
answer:
coord: [2, -2]
final_pos: [16, 3]
final_ori: [1, 0]
room_id: 2
object_positions:
bike: [9, 6]
airplane: [15, 2]
rubberduck: [1, 6]
lego man: [9, 8]
toilet: [19, 4]
truck: [5, 6]
backpack: [11, 3]
shoppingcart: [13, 3]
fridge: [17, 2]
soda can: [5, 8]
basket: [17, 6]
lamp: [1, 10]
red door: [7, 5]
green door: [14, 5]
object_orientations:
bike: [0, -1]
airplane: [0, -1]
rubberduck: [0, 1]
lego man: [0, -1]
toilet: [1, 0]
truck: [1, 0]
backpack: [0, -1]
shoppingcart: [-1, 0]
red door: [1, 0]
green door: [1, 0]
final_observation: [{'name': 'toilet', 'direction': 'front-slight-left', 'distance': 'mid distance', 'orientation': 'facing forward'}, {'name': 'fridge', 'direction': 'front-right', 'distance': 'near', 'orientation': None}]
gate_info:
red door:
room_ids: [3, 1]
ori_by_room:
3: [-1, 0]
1: [1, 0]
green door:
room_ids: [1, 2]
ori_by_room:
1: [-1, 0]
2: [1, 0]
object_rooms:
bike: 1
airplane: 2
rubberduck: 3
lego man: 1
toilet: 2
truck: 3
backpack: 1
shoppingcart: 1
fridge: 2
soda can: 3
basket: 2
lamp: 3
red door: [3, 1]
green door: [1, 2]
task_type: View2LocationTextEvaluationTask
choices: []
kwargs:
(none)
similarity: 1
raw_score: 1
score: 1
evaluation_mode: use_model_cogmap
Evaluation state
Q1: View2LocationTextEvaluationTask

Question 2 (ID: 9d51c9ea1c03280c)

✅ Evaluation Results
id: 9d51c9ea1c03280c
question: You move to a new location and face north. You observe: basket is at front-slight-right, mid distance; green door is at front-left, mid distance, on left wall Treat the green door as the new 'origin' (0, 0). What is your new 2D coordinate (x, y)? Answer format: (x, y) Example: (2, -1)
action: None
answer:
coord: [2, -2]
final_pos: [16, 3]
final_ori: [0, 1]
room_id: 2
object_positions:
bike: [9, 6]
airplane: [15, 2]
rubberduck: [1, 6]
lego man: [9, 8]
toilet: [19, 4]
truck: [5, 6]
backpack: [11, 3]
shoppingcart: [13, 3]
fridge: [17, 2]
soda can: [5, 8]
basket: [17, 6]
lamp: [1, 10]
red door: [7, 5]
green door: [14, 5]
object_orientations:
bike: [0, -1]
airplane: [0, -1]
rubberduck: [0, 1]
lego man: [0, -1]
toilet: [1, 0]
truck: [1, 0]
backpack: [0, -1]
shoppingcart: [-1, 0]
red door: [1, 0]
green door: [1, 0]
final_observation: [{'name': 'basket', 'direction': 'front-slight-right', 'distance': 'mid distance', 'orientation': None}, {'name': 'green door', 'direction': 'front-left', 'distance': 'mid distance', 'orientation': 'on left wall'}]
gate_info:
red door:
room_ids: [3, 1]
ori_by_room:
3: [-1, 0]
1: [1, 0]
green door:
room_ids: [1, 2]
ori_by_room:
1: [-1, 0]
2: [1, 0]
object_rooms:
bike: 1
airplane: 2
rubberduck: 3
lego man: 1
toilet: 2
truck: 3
backpack: 1
shoppingcart: 1
fridge: 2
soda can: 3
basket: 2
lamp: 3
red door: [3, 1]
green door: [1, 2]
task_type: View2LocationTextEvaluationTask
choices: []
kwargs:
(none)
similarity: 1
raw_score: 1
score: 1
evaluation_mode: use_model_cogmap
Evaluation state
Q2: View2LocationTextEvaluationTask

Question 3 (ID: 231594f7927a5d40)

✅ Evaluation Results
id: 231594f7927a5d40
question: You move to a new location and face south. You observe: truck is at front-slight-left, mid distance, facing left; rubberduck is at front-right, slightly far, facing backward; red door is at front-left, slightly far, on left wall Treat the red door as the new 'origin' (0, 0). What is your new 2D coordinate (x, y)? Answer format: (x, y) Example: (2, -1)
action: None
answer:
coord: [-3, 4]
final_pos: [4, 9]
final_ori: [0, -1]
room_id: 3
object_positions:
bike: [9, 6]
airplane: [15, 2]
rubberduck: [1, 6]
lego man: [9, 8]
toilet: [19, 4]
truck: [5, 6]
backpack: [11, 3]
shoppingcart: [13, 3]
fridge: [17, 2]
soda can: [5, 8]
basket: [17, 6]
lamp: [1, 10]
red door: [7, 5]
green door: [14, 5]
object_orientations:
bike: [0, -1]
airplane: [0, -1]
rubberduck: [0, 1]
lego man: [0, -1]
toilet: [1, 0]
truck: [1, 0]
backpack: [0, -1]
shoppingcart: [-1, 0]
red door: [1, 0]
green door: [1, 0]
final_observation: [{'name': 'truck', 'direction': 'front-slight-left', 'distance': 'mid distance', 'orientation': 'facing left'}, {'name': 'rubberduck', 'direction': 'front-right', 'distance': 'slightly far', 'orientation': 'facing backward'}, {'name': 'red door', 'direction': 'front-left', 'distance': 'slightly far', 'orientation': 'on left wall'}]
gate_info:
red door:
room_ids: [3, 1]
ori_by_room:
3: [-1, 0]
1: [1, 0]
green door:
room_ids: [1, 2]
ori_by_room:
1: [-1, 0]
2: [1, 0]
object_rooms:
bike: 1
airplane: 2
rubberduck: 3
lego man: 1
toilet: 2
truck: 3
backpack: 1
shoppingcart: 1
fridge: 2
soda can: 3
basket: 2
lamp: 3
red door: [3, 1]
green door: [1, 2]
task_type: View2LocationTextEvaluationTask
choices: []
kwargs:
(none)
similarity: 0.819
raw_score: 0.819
score: 0.819
evaluation_mode: use_model_cogmap
Evaluation state
Q3: View2LocationTextEvaluationTask

📊 Task: DirectionEvaluationTask

Question 1 (ID: 9d13d2040c8f0e75)

✅ Evaluation Results
id: 9d13d2040c8f0e75
question: You return to your starting position and face north. From a Top-Down map, describe where basket is relative to toilet. Answer format: <cardinal direction>, <distance> Example: north-west, near
action: None
answer: north west, mid distance
task_type: DirectionEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: use_model_cogmap
Evaluation state
Q1: DirectionEvaluationTask

Question 2 (ID: e2300d26c5895d2d)

✅ Evaluation Results
id: e2300d26c5895d2d
question: You return to your starting position and face north. From a Top-Down map, describe where rubberduck is relative to lego man. Answer format: <cardinal direction>, <distance> Example: north-west, near
action: None
answer: west, far
task_type: DirectionEvaluationTask
choices: []
kwargs:
(none)
score: 0.500
evaluation_mode: use_model_cogmap
Evaluation state
Q2: DirectionEvaluationTask

Question 3 (ID: c2a21d8364e4dca4)

✅ Evaluation Results
id: c2a21d8364e4dca4
question: You return to your starting position and face north. From a Top-Down map, describe where shoppingcart is relative to airplane. Answer format: <cardinal direction>, <distance> Example: north-west, near
action: None
answer: north west, mid distance
task_type: DirectionEvaluationTask
choices: []
kwargs:
(none)
score: 0
evaluation_mode: use_model_cogmap
Evaluation state
Q3: DirectionEvaluationTask

📊 Task: PovEvaluationTask

Question 1 (ID: 063960e28e84e1b8)

✅ Evaluation Results
id: 063960e28e84e1b8
question: Now you jump to rubberduck's direction, facing its direction. Describe where lamp is relative to you. Answer format: <ego direction>, <distance> Example: front-left, near
action: None
answer: front, mid distance
task_type: PovEvaluationTask
choices: []
kwargs:
(none)
score: 0.500
evaluation_mode: use_model_cogmap
Evaluation state
Q1: PovEvaluationTask

Question 2 (ID: dbbb7b93fba9d779)

✅ Evaluation Results
id: dbbb7b93fba9d779
question: Now you jump to lego man's direction, facing its direction. Describe where shoppingcart is relative to you. Answer format: <ego direction>, <distance> Example: front-left, near
action: None
answer: front-left, slightly far
task_type: PovEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: use_model_cogmap
Evaluation state
Q2: PovEvaluationTask

Question 3 (ID: b2e43f1e4b97a13e)

✅ Evaluation Results
id: b2e43f1e4b97a13e
question: Now you jump to shoppingcart's direction, facing its direction. Describe where bike is relative to you. Answer format: <ego direction>, <distance> Example: front-left, near
action: None
answer: front-right, slightly far
task_type: PovEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: use_model_cogmap
Evaluation state
Q3: PovEvaluationTask

📊 Task: BackwardPovTextEvaluationTask

Question 1 (ID: 4022a79578213eb8)

✅ Evaluation Results
id: 4022a79578213eb8
question: Now you jump to an object's position, facing its direction. You observe that shoppingcart is front-left, slightly far, facing right; red door is front-right, mid distance, on right wall; backpack is front-slight-left, slightly far, facing forward. Which object are you standing at? Answer format: <object_name> Example: lamp
action: None
answer:
answer: lego man
final_pos: [9, 8]
final_ori: [0, -1]
final_observation: [{'name': 'shoppingcart', 'direction': 'front-left', 'distance': 'slightly far', 'orientation': 'facing right'}, {'name': 'red door', 'direction': 'front-right', 'distance': 'mid distance', 'orientation': 'on right wall'}, {'name': 'backpack', 'direction': 'front-slight-left', 'distance': 'slightly far', 'orientation': 'facing forward'}]
object_positions:
bike: [9.0, 6.0]
airplane: [15.0, 2.0]
rubberduck: [1.0, 6.0]
lego man: [9.0, 8.0]
toilet: [19.0, 4.0]
truck: [5.0, 6.0]
backpack: [11.0, 3.0]
shoppingcart: [13.0, 3.0]
fridge: [17.0, 2.0]
soda can: [5.0, 8.0]
basket: [17.0, 6.0]
lamp: [1.0, 10.0]
red door: [7.0, 5.0]
green door: [14.0, 5.0]
object_orientations:
bike: [0, -1]
airplane: [0, -1]
rubberduck: [0, 1]
lego man: [0, -1]
toilet: [1, 0]
truck: [1, 0]
backpack: [0, -1]
shoppingcart: [-1, 0]
red door: [1, 0]
green door: [1, 0]
room_id: 1
gate_info:
red door:
room_ids: [3, 1]
ori_by_room:
3: [-1, 0]
1: [1, 0]
green door:
room_ids: [1, 2]
ori_by_room:
1: [-1, 0]
2: [1, 0]
object_rooms:
bike: 1
airplane: 2
rubberduck: 3
lego man: 1
toilet: 2
truck: 3
backpack: 1
shoppingcart: 1
fridge: 2
soda can: 3
basket: 2
lamp: 3
red door: [3, 1]
green door: [1, 2]
task_type: BackwardPovTextEvaluationTask
choices: []
kwargs:
(none)
name_match: True
view_match: True
score: 1
evaluation_mode: use_model_cogmap
Evaluation state
Q1: BackwardPovTextEvaluationTask

Question 2 (ID: d217d8691521707e)

✅ Evaluation Results
id: d217d8691521707e
question: Now you jump to an object's position, facing its direction. You observe that red door is front-right, mid distance, on front wall. Which object are you standing at? Answer format: <object_name> Example: lamp
action: None
answer:
answer: truck
final_pos: [5, 6]
final_ori: [1, 0]
final_observation: [{'name': 'red door', 'direction': 'front-right', 'distance': 'mid distance', 'orientation': 'on front wall'}]
object_positions:
bike: [9.0, 6.0]
airplane: [15.0, 2.0]
rubberduck: [1.0, 6.0]
lego man: [9.0, 8.0]
toilet: [19.0, 4.0]
truck: [5.0, 6.0]
backpack: [11.0, 3.0]
shoppingcart: [13.0, 3.0]
fridge: [17.0, 2.0]
soda can: [5.0, 8.0]
basket: [17.0, 6.0]
lamp: [1.0, 10.0]
red door: [7.0, 5.0]
green door: [14.0, 5.0]
object_orientations:
bike: [0, -1]
airplane: [0, -1]
rubberduck: [0, 1]
lego man: [0, -1]
toilet: [1, 0]
truck: [1, 0]
backpack: [0, -1]
shoppingcart: [-1, 0]
red door: [1, 0]
green door: [1, 0]
room_id: 3
gate_info:
red door:
room_ids: [3, 1]
ori_by_room:
3: [-1, 0]
1: [1, 0]
green door:
room_ids: [1, 2]
ori_by_room:
1: [-1, 0]
2: [1, 0]
object_rooms:
bike: 1
airplane: 2
rubberduck: 3
lego man: 1
toilet: 2
truck: 3
backpack: 1
shoppingcart: 1
fridge: 2
soda can: 3
basket: 2
lamp: 3
red door: [3, 1]
green door: [1, 2]
task_type: BackwardPovTextEvaluationTask
choices: []
kwargs:
(none)
name_match: True
view_match: True
score: 1
evaluation_mode: use_model_cogmap
Evaluation state
Q2: BackwardPovTextEvaluationTask

Question 3 (ID: 7599c12dc388ac1a)

✅ Evaluation Results
id: 7599c12dc388ac1a
question: Now you jump to an object's position, facing its direction. You observe that lamp is front, mid distance. Which object are you standing at? Answer format: <object_name> Example: lamp
action: None
answer:
answer: rubberduck
final_pos: [1, 6]
final_ori: [0, 1]
final_observation: [{'name': 'lamp', 'direction': 'front', 'distance': 'mid distance', 'orientation': None}]
object_positions:
bike: [9.0, 6.0]
airplane: [15.0, 2.0]
rubberduck: [1.0, 6.0]
lego man: [9.0, 8.0]
toilet: [19.0, 4.0]
truck: [5.0, 6.0]
backpack: [11.0, 3.0]
shoppingcart: [13.0, 3.0]
fridge: [17.0, 2.0]
soda can: [5.0, 8.0]
basket: [17.0, 6.0]
lamp: [1.0, 10.0]
red door: [7.0, 5.0]
green door: [14.0, 5.0]
object_orientations:
bike: [0, -1]
airplane: [0, -1]
rubberduck: [0, 1]
lego man: [0, -1]
toilet: [1, 0]
truck: [1, 0]
backpack: [0, -1]
shoppingcart: [-1, 0]
red door: [1, 0]
green door: [1, 0]
room_id: 3
gate_info:
red door:
room_ids: [3, 1]
ori_by_room:
3: [-1, 0]
1: [1, 0]
green door:
room_ids: [1, 2]
ori_by_room:
1: [-1, 0]
2: [1, 0]
object_rooms:
bike: 1
airplane: 2
rubberduck: 3
lego man: 1
toilet: 2
truck: 3
backpack: 1
shoppingcart: 1
fridge: 2
soda can: 3
basket: 2
lamp: 3
red door: [3, 1]
green door: [1, 2]
task_type: BackwardPovTextEvaluationTask
choices: []
kwargs:
(none)
name_match: True
view_match: True
score: 1
evaluation_mode: use_model_cogmap
Evaluation state
Q3: BackwardPovTextEvaluationTask

📊 Task: Action2ViewEvaluationTask

Question 1 (ID: e5146a0f564c765f)

✅ Evaluation Results
id: e5146a0f564c765f
question: You return to your starting position and face north. You will execute the following action sequence: 1. Rotate(90) 2. Jump to the object at front, near. 3. Jump to the object at front-right, slightly far. After executing the actions, what is the ego relation of toilet relative to you? Answer format: <ego direction>, <distance> Example: front, near
action: None
answer: front-left, mid distance
task_type: Action2ViewEvaluationTask
choices: []
kwargs:
(none)
score: 0.500
evaluation_mode: use_model_cogmap
Evaluation state
Q1: Action2ViewEvaluationTask

Question 2 (ID: 90f3cb1a03f64a70)

✅ Evaluation Results
id: 90f3cb1a03f64a70
question: You return to your starting position and face north. You will execute the following action sequence: 1. Rotate(90) 2. Jump to the object at front, near. 3. Jump to the object at front-right, slightly far. 4. Rotate(-180) After executing the actions, what is the ego relation of green door relative to you? Answer format: <ego direction>, <distance> Example: front, near
action: None
answer: front-right, slightly far
task_type: Action2ViewEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: use_model_cogmap
Evaluation state
Q2: Action2ViewEvaluationTask

Question 3 (ID: 7b132df57af38164)

✅ Evaluation Results
id: 7b132df57af38164
question: You return to your starting position and face north. You will execute the following action sequence: 1. Jump to the object at front-left, slightly far. 2. Rotate(90) 3. Jump to the object at front-right, slightly far. 4. Rotate(90) After executing the actions, what is the ego relation of fridge relative to you? Answer format: <ego direction>, <distance> Example: front, near
action: None
answer: front-left, slightly far
task_type: Action2ViewEvaluationTask
choices: []
kwargs:
(none)
score: 1
evaluation_mode: use_model_cogmap
Evaluation state
Q3: Action2ViewEvaluationTask