Mind the Perspective: Let's Reason Recursively for Theory of Mind
Quick Answer
RecToM introduces a recursive perspective construction framework for Theory of Mind (ToM) reasoning, outperforming advanced models like GPT-5.4 and Qwen3.5 with 100% accuracy on the Hi-ToM benchmark.
Quick Take
This method effectively models nested beliefs, addressing challenges in inferring agents' beliefs from limited observations.
Key Points
- RecToM models nested beliefs through recursive perspective construction.
- Achieves state-of-the-art performance on ToM benchmarks like Hi-ToM and Big-ToM.
- Outperforms recent advanced approaches consistently across multiple backbones.
- KD45 analysis shows well-formed belief modality beyond simple event filtering.
- 100% accuracy on Hi-ToM with GPT-5.4 and Qwen3.5 demonstrates effectiveness.
Paper Resources
Article Content
From source RSS / original summaryarXiv:2606. 11724v1 Announce Type: new Abstract: Theory of Mind (ToM) reasoning requires inferring agents' beliefs from partial and asymmetric observations, which remains an open challenge for . Existing prompting-based approaches improve ToM reasoning through observable-event filtering or temporal belief chains, without explicitly modeling nested beliefs. We introduce RecToM, an inference-time framework for ToM reasoning that models nested beliefs via recursive perspective construction.
RecToM constructs each character perspective from the preceding character perspective along the character chain specified by the question, reducing higher-order belief questions to actual-world questions within the final constructed perspective. We further provide a KD45 analysis showing that RecToM's perspective construction induces a well-formed belief modality beyond simple event filtering.
Experiments on ToM benchmarks, including Hi-ToM, Big-ToM, and FanToM, across multiple LLM backbones show that RecToM consistently outperforms recent advanced approaches, achieving state-of-the-art performance. Notably, RecToM reaches 100\% accuracy on Hi-ToM with GPT-5. 4 and Qwen3. 5, a benchmark requiring higher-order ToM reasoning.
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.AI
See more →HOBA: Hierarchical On-Policy Bidding Agents for Adaptive Online Advertising
HOBA (Hierarchical On-policy Bidding Agents) is a novel hierarchical reinforcement learning framework that enhances online advertising bidding systems by improving adaptability and reducing hyperparameter tuning costs. It utilizes a for hyperparameter inference, a SARSA agent for expert model selection, and a dynamic expert pool for bid execution, achieving a +3.6% increase in target cost during large-scale deployment and outperforming state-of-the-art baselines on AuctionNet.