Research map / Large language models / Agents, code and math
MovieGraph-ToM: Evaluating Long-Range Theory of Mind in Large Language Models via Implicit Social-Causal Graphs
Tingjiang Wei, Qin Ni, Rong Gao, Yingying Wang, Liang He
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| switch from single-choice to multiple-choice questions | gpt-4o-mini · MovieGraph-ToM | accuracy: 93.8 -> 35.4 | Hurt |
| switch from single-choice to multiple-choice questions | llava-onevision-qwen2-7b · MovieGraph-ToM | accuracy: 84.6 -> 26.9 | Hurt |
| switch from single-choice to multiple-choice questions | gemini-2.5-pro · MovieGraph-ToM | accuracy: 96.4 -> 70.8 | Hurt |
| switch from multiple-choice to open-ended generation | gpt-4.1 and gemini-2.5-pro · MovieGraph-ToM | open-ended score: 0.675 and 0.674 (below 0.7 threshold) | Hurt |
| increase reasoning depth from 1 to 3 | gemini-2.5-pro · MovieGraph-ToM | explanation score: 0.892 -> 0.760 | Hurt |
| increase reasoning depth to 3 | most models · MovieGraph-ToM | accuracy | No clear effect |
| use larger / stronger model | gemini-2.5-pro vs smaller models · MovieGraph-ToM | overall accuracy: 75.1 -> 92.0 | Improved |
| add counterfactual premise to unexplored branches | all evaluated models · MovieGraph-ToM | Counterfactual Recovery Score | Improved |
Closest papers by meaning
- OmniToM: Benchmarking Theory of Mind in LLMs via Explicit Belief Modeling (NeurIPS 2026 · similarity 0.91)
- Language Models Represent Beliefs of Self and Others (ICML 2024 · similarity 0.91)
- MetaMind: Modeling Human Social Thoughts with Metacognitive Multi-Agent Systems (NeurIPS 2025 · similarity 0.90)
- Video-Only ToM: Enhancing Theory of Mind in Multimodal Large Language Models (CVPR 2026 · similarity 0.90)
- Theory of Mind and Persuasion Beyond Conversation: Assessing the Capacity of LLMs to Induce Belief States via Planning and Action (NeurIPS 2026 · similarity 0.89)
- SoMi-ToM: Evaluating Multi-Perspective Theory of Mind in Embodied Social Interactions (NeurIPS 2025 · similarity 0.89)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
