ToM2Vec: Learning multi-agent theory of mind with continuous reasoning tokens
What happened
Nithin Sivakumaran and colleagues propose ToM2Vec, a method that lets a model form a theory of mind (inferring what its partner is thinking) through continuous reasoning tokens under a fixed budget, using partners' actual thoughts as training supervision. The authors say the method delivers consistent gains over the base model and several competitive baselines in cooperative, competitive, and mixed-motive environments, including Overcooked-AI, multi-party negotiation, Decrypto, Hanabi, and Collab Escape/Capture. The current basis is mainly the abstract of this arXiv paper, and its performance claims come from the authors' own reports. No independent replication or third-party evaluation has been seen yet.
AI-written from the coverage · updated 2026-10-22
Coverage timeline
Follow the coverage to see the story from different sides.
- arXiv Board & Classic Game AIPickToM2Vec: Continuous theory-of-mind tokens for multi-agent coordination
ToM2Vec learns theory of mind for multi-agent coordination through a fixed budget of continuous reasoning tokens, using partners' actual thoughts as training supervision. Across cooperative, competitive, and mixed-motive environments, including Overcooked-AI, multi-party negotiation, Decrypto, Hanabi, and Collab Escape/Capture, it delivers consistent gains over the base model and several competitive baselines.