Distractor Generation on Discrete_40
3.25PlausibilityMCTS-guided reasoning reconstruction framework
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| MCTS-guided reasoning reconstruction frameworkBackbone Model=Claude-4-Sonnet2025.08 | 3.25 | 2.64 | |
| MCTS-guided reasoning reconstruction frameworkBackbone Model=Deepseek-V32025.08 | 3.16 | 2.59 | |
| CoTBackbone Model=Deepseek-V32025.08 | 2.89 | 2.27 | |
| MCTS-guided reasoning reconstruction frameworkBackbone Model=GPT-4o2025.08 | 2.88 | 2.43 | |
| CoTBackbone Model=Claude-4-Sonnet2025.08 | 2.85 | 2.39 | |
| MCTS-guided reasoning reconstruction frameworkBackbone Model=GPT-3.5-turbo2025.08 | 2.81 | 2.23 | |
| MCTS-guided reasoning reconstruction frameworkBackbone Model=LLaMA-3-70B2025.08 | 2.77 | 2.32 | |
| CoTBackbone Model=GPT-4o2025.08 | 2.72 | 2.25 | |
| CoTBackbone Model=GPT-3.5-turbo2025.08 | 2.64 | 2.12 | |
| CoTBackbone Model=LLaMA-3-70B2025.08 | 2.54 | 2.13 |