List your company

Group-level distractor generation on CEval Discrete Math

45.56RecallClaude-4-Sonnet
17.823225.024132.22539.4259Aug 15, 2025

Evaluation results

MethodLinks
Claude-4-SonnetFramework=MCTS-guided reasoning reconstruction, Backbone=Claude-4-Sonnet2025.0845.56
GPT-4oFramework=MCTS-guided reasoning reconstruction, Backbone=GPT-4o2025.0845.56
GPT-3.5-turboFramework=MCTS-guided reasoning reconstruction, Backbone=GPT-3.5-turbo2025.0844.44
Deepseek-V3Framework=MCTS-guided reasoning reconstruction, Backbone=Deepseek-V32025.0840
LLaMA-3-70BFramework=MCTS-guided reasoning reconstruction, Backbone=LLaMA-3-70B2025.0836.67
D-GEN2025.0828.89
DiVERT2025.0820
O&A2025.0818.89