ToM2Vec encodes Theory of Mind as vectors, beating baselines in Overcooked agent QA
EliasEskin · x · 2026-10-22
New research from UNC, Cisco, and UT Austin introduces ToM2Vec, which encodes Theory of Mind into vectors for multi-agent collaboration.
- In Overcooked, agents were probed with QA about their partner's beliefs, desires, and next actions (Bayesian ToM properties) to test what the vectors encode.
- ToM2Vec achieved the highest accuracy across all three categories, beating each strongest baseline by 5.6, 2.0, and 2.7 percentage points.
- Paper, HuggingFace models, and code are all publicly available.
Related event: ToM2Vec: Continuous Theory-of-Mind vectors help LLM agents model teammates(5 posts)→
More from Research
- Sheffield advertises funded PhD on reliability and bias of LLM-based evaluation — nikaletras · 2026-10-22
- Scientific bottlenecks: reinvent assays, don't just automate them — owl_posting · 2026-10-22
- AI academic writing's vague-language trap: it reads as consistent to you but confuses everyone else — anshulkundaje · 2026-10-22
- YC hosts AI for Biology Paper Club as 81% of US doctors adopt AI — ycombinator · 2026-10-22
- CONCERT: Generative Modeling of Spatially Resolved Perturbation Responses — anshulkundaje · 2026-10-22
- ToM2Vec trained via self-play with Qwen3-8B transfers to GPT-5.4 teams, adding 17.3 points — EliasEskin · 2026-10-22