Reasoning
Embodied Web Agents
Embodied Web Agents (indoor cooking, text-based): per-task pass/fail for the GPT (gpt-4o) run. Each task pairs a web subtask (find/buy recipe items) with an embodied subtask (achieve the recipe object final states in AI2-THOR). Released as two per-task score files.
876items
1subjects
MITlicense
agents_and_tool_usedomain
reasoningdomain
textmodality
item-level responses released
Saturation status: Unknown
Response matrix
Fit to width. Hover for subject & item; click a cell for details.
Correct (1)Incorrect (0)Unobserved
Scale: 1 = correct · 0 = incorrect