3260 papers • 126 benchmarks • 313 datasets
This task has no description! Would you like to contribute one?
(Image credit: Papersgraph)
These leaderboards are used to track progress in theory-of-mind-modeling-2
No benchmarks available.
Use these libraries to find theory-of-mind-modeling-2 models and implementations
No subtasks available.
A fine-grained dataset of collaborative tasks performed by pairs of human subjects in the 3D virtual blocks world of Minecraft is introduced to enable theory of mind modeling in situated interactions.
It is found that humans struggle to fully integrate information from teammates into their decisions, especially when communication load is high, and have cognitive biases which lead them to underweight certain useful, but ambiguous, information.
A novel problem for agents to predict the missing task knowledge for themselves and for their partners based on rich perceptual and dialogue history is formulated and empirical results suggest that predicting the partner's missing knowledge is a more viable approach than predicting one's own.
This position paper argues for a holistic and situated evaluation of ToM to break ToM into individual components and treat LLMs as an agent who is physically situated in environments and socially situated in interactions with humans.
A multimodal Theory of Mind question answering (MMToM-QA) benchmark is introduced and a systematic comparison of human performance, BIP-ALM, and state-of-the-art models, including GPT-4 are compared.
Adding a benchmark result helps the community track progress.