Combinatorial Reasoning on Graph Coloring
100AccuracyBoN
Evaluation Results
| Method | Links | |
|---|---|---|
| BoNAlgorithm=BoN, Model=Llama 70B2025.03 | 100 | |
| SEM-CTRLAlgorithm=SEM-CTRL, Model=Llama 1B2025.03 | 100 | |
| SEM-CTRLAlgorithm=SEM-CTRL, Model=Llama 70B2025.03 | 100 | |
| SoftCoTBase model=Qwen2.5, Transfer cost=full retrain2026.05 | 79 | |
| o1-previewAlgorithm=API, Model=o1-preview2025.03 | 75 | |
| DeepSeek-R1Algorithm=API, Model=DeepSeek-R12025.03 | 75 | |
| o4-miniAlgorithm=API, Model=o4-mini2025.03 | 75 | |
| SoftCoTBase model=GPT-OSS, Transfer cost=full retrain2026.05 | 68 | |
| PT-SFTBase model=Qwen2.5, Transfer cost=full retrain2026.05 | 64 | |
| HyperGuideBase model=Qwen2.5, Transfer cost=small MLP2026.05 | 64 | |
| Few-shotBase model=Qwen2.5, Transfer cost=N/A2026.05 | 63 | |
| HyperGuideBase model=GPT-OSS, Transfer cost=small MLP2026.05 | 63 | |
| Self-ConsistencyBase model=Qwen2.5, Transfer cost=N/A2026.05 | 60 | |
| OVMBase model=Qwen2.5, Transfer cost=full retrain2026.05 | 59.4 | |
| PT-SFTBase model=GPT-OSS, Transfer cost=full retrain2026.05 | 58 | |
| Self-ConsistencyBase model=GPT-OSS, Transfer cost=N/A2026.05 | 57.4 | |
| SoftCoTBase model=Mistral, Transfer cost=full retrain2026.05 | 57 | |
| HyperGuideBase model=Mistral, Transfer cost=small MLP2026.05 | 57 | |
| OVMBase model=Mistral, Transfer cost=full retrain2026.05 | 55.4 | |
| OVMBase model=GPT-OSS, Transfer cost=full retrain2026.05 | 53 | |
| PT-SFTBase model=Mistral, Transfer cost=full retrain2026.05 | 52 | |
| Few-shotBase model=GPT-OSS, Transfer cost=N/A2026.05 | 51 | |
| Self-ConsistencyBase model=Mistral, Transfer cost=N/A2026.05 | 49.6 | |
| Tree of ThoughtsBase model=GPT-OSS, Transfer cost=N/A2026.05 | 49 | |
| Few-shotBase model=Mistral, Transfer cost=N/A2026.05 | 49 | |
| BaseAlgorithm=Base, Model=Llama 70B2025.03 | 37.5 | |
| Tree of ThoughtsBase model=Qwen2.5, Transfer cost=N/A2026.05 | 34 | |
| Tree of ThoughtsBase model=Mistral, Transfer cost=N/A2026.05 | 14 | |
| BaseAlgorithm=Base, Model=Llama 1B2025.03 | 0 | |
| BoNAlgorithm=BoN, Model=Llama 1B2025.03 | 0 |