List your company

Scientific Question Answering on SciEval Static

81.52AccuracyReason, Reward, Refine
43.029653.022363.01573.0077Jul 6, 2026

Evaluation results

MethodLinks
Reason, Reward, RefineModel=LLaMA 3.2 3B2026.0781.52
Reason, Reward, RefineModel=Qwen 2.5 1.5B2026.0779.29
RAGModel=Qwen 2.5 1.5B2026.0768.12
Reason, Reward, RefineModel=LLaMA 3.2 1B2026.0768.09
RAGModel=LLaMA 3.2 3B2026.0766.87
Reason, Reward, RefineModel=Phi 3.5 Mini 3.8B2026.0765.19
CoTModel=Qwen 2.5 1.5B2026.0762.73
CoTModel=Phi 3.5 Mini 3.8B2026.0762.43
CoTModel=LLaMA 3.2 3B2026.0762.26
SFTModel=LLaMA 3.2 3B2026.0761.59
RAGModel=Phi 3.5 Mini 3.8B2026.0761.01
DPOModel=Qwen 2.5 1.5B2026.0759.32
SFTModel=Phi 3.5 Mini 3.8B2026.0759.17
SFTModel=Qwen 2.5 1.5B2026.0757.93
DPOModel=LLaMA 3.2 3B2026.0757.93
RAGModel=LLaMA 3.2 1B2026.0757.37
DPOModel=LLaMA 3.2 1B2026.0756.93
DPOModel=Phi 3.5 Mini 3.8B2026.0756.87
SFTModel=LLaMA 3.2 1B2026.0755.49
CoTModel=LLaMA 3.2 1B2026.0744.51