Belief Prediction on ToM-BPD (test)
54.64Belief AccuracyTTBYS
Evaluation Results
| Method | Links | |
|---|---|---|
| TTBYSBackbone=Qwen-3-8B, Enhancement=TTBYS2026.05 | 54.64 | |
| TTBYSBackbone=LLaMA-3.1-8B-Instruct, Enhancement=TTBYS2026.05 | 43.62 | |
| Qwen-3-8B + MMBackbone=Qwen-3-8B, Enhancement=MM2026.05 | 42.05 | |
| Qwen-3-8B + HMBackbone=Qwen-3-8B, Enhancement=HM2026.05 | 41.21 | |
| Qwen-3-8B + CoTBackbone=Qwen-3-8B, Enhancement=CoT2026.05 | 39.42 | |
| Qwen-3-8BBackbone=Qwen-3-8B, Enhancement=Basic2026.05 | 38.64 | |
| TTBYSBackbone=Mixtral-7B-Instruct, Enhancement=TTBYS2026.05 | 36.96 | |
| GPT-4o-mini + MMBackbone=GPT-4o-mini, Enhancement=MM2026.05 | 31.96 | |
| GPT-5 + MMBackbone=GPT-5, Enhancement=MM2026.05 | 31.74 | |
| LLaMA-3.1-8B-Instruct + MMBackbone=LLaMA-3.1-8B-Instruct, Enhancement=MM2026.05 | 31.67 | |
| LLaMA-3.1-8B-Instruct + HMBackbone=LLaMA-3.1-8B-Instruct, Enhancement=HM2026.05 | 30.8 | |
| GPT-4o-mini + HMBackbone=GPT-4o-mini, Enhancement=HM2026.05 | 30.6 | |
| GPT-5 + HMBackbone=GPT-5, Enhancement=HM2026.05 | 29.6 | |
| Gemini-3-pro-high + MMBackbone=Gemini-3-pro-high, Enhancement=MM2026.05 | 29.28 | |
| Mixtral-7B-Instruct + MMBackbone=Mixtral-7B-Instruct, Enhancement=MM2026.05 | 29.06 | |
| Gemini-3-pro-high + HMBackbone=Gemini-3-pro-high, Enhancement=HM2026.05 | 27.9 | |
| LLaMA-3.1-8B-Instruct + CoTBackbone=LLaMA-3.1-8B-Instruct, Enhancement=CoT2026.05 | 27.69 | |
| Mixtral-7B-Instruct + HMBackbone=Mixtral-7B-Instruct, Enhancement=HM2026.05 | 26.8 | |
| LLaMA-3.1-8B-InstructBackbone=LLaMA-3.1-8B-Instruct, Enhancement=Basic2026.05 | 25.67 | |
| Mixtral-7B-Instruct + CoTBackbone=Mixtral-7B-Instruct, Enhancement=CoT2026.05 | 24.13 | |
| GPT-4o-mini + CoTBackbone=GPT-4o-mini, Enhancement=CoT2026.05 | 23.41 | |
| Mixtral-7B-InstructBackbone=Mixtral-7B-Instruct, Enhancement=Basic2026.05 | 22.86 | |
| GPT-5 + CoTBackbone=GPT-5, Enhancement=CoT2026.05 | 22.68 | |
| GPT-5Backbone=GPT-5, Enhancement=Basic2026.05 | 21.84 | |
| GPT-4o-miniBackbone=GPT-4o-mini, Enhancement=Basic2026.05 | 21.45 | |
| Gemini-3-pro-high + CoTBackbone=Gemini-3-pro-high, Enhancement=CoT2026.05 | 18.34 | |
| Gemini-3-pro-highBackbone=Gemini-3-pro-high, Enhancement=Basic2026.05 | 17.84 |