Logical Reasoning on Rule-chaining (test)
84AccuracyOVM
Evaluation results
| Method | Links | |
|---|---|---|
| OVMBase model=Qwen2.52026.05 | 84 | |
| HyperGuideBase model=Qwen2.52026.05 | 80 | |
| HyperGuideBase model=Mistral2026.05 | 79 | |
| Self-ConsistencyBase model=Qwen2.52026.05 | 78 | |
| PT-SFTBase model=Qwen2.52026.05 | 77 | |
| HyperGuideBase model=GPT-OSS2026.05 | 77 | |
| Self-ConsistencyBase model=Mistral2026.05 | 74.5 | |
| SoftCoTBase model=Qwen2.52026.05 | 73.5 | |
| PT-SFTBase model=GPT-OSS2026.05 | 72 | |
| PT-SFTBase model=Mistral2026.05 | 72 | |
| OVMBase model=Mistral2026.05 | 71.5 | |
| OVMBase model=GPT-OSS2026.05 | 70 | |
| SoftCoTBase model=GPT-OSS2026.05 | 67.5 | |
| SoftCoTBase model=Mistral2026.05 | 63.5 | |
| Few-shotBase model=Mistral2026.05 | 54 | |
| Few-shotBase model=Qwen2.52026.05 | 53 | |
| Tree of ThoughtsBase model=Qwen2.52026.05 | 52 | |
| Self-ConsistencyBase model=GPT-OSS2026.05 | 52 | |
| Tree of ThoughtsBase model=GPT-OSS2026.05 | 50 | |
| Tree of ThoughtsBase model=Mistral2026.05 | 50 | |
| Few-shotBase model=GPT-OSS2026.05 | 16.1 |