List your company

Free-form answer generation on KLAR (train)

49.69AccuracyGRPO
15.692424.518733.34542.1713Jun 4, 2026

Evaluation results

MethodLinks
GRPOBackbone=Qwen-2.5-7B2026.0649.69
BaselineBackbone=Qwen-2.5-7B2026.0647.68
CPT + GRPOBackbone=Qwen-2.5-7B2026.0645.97
CPT + SFTBackbone=Qwen-2.5-7B2026.0645.43
SFTBackbone=Qwen-2.5-7B2026.0643.23
CPTBackbone=Qwen-2.5-7B2026.0638.99
CPT + GRPOBackbone=OLMo-2-1124-7B2026.0629.8
GRPOBackbone=OLMo-2-1124-7B2026.0629
BaselineBackbone=OLMo-2-1124-7B2026.0624.6
CPT + SFTBackbone=OLMo-2-1124-7B2026.0620.7
SFTBackbone=OLMo-2-1124-7B2026.0618.1
CPTBackbone=OLMo-2-1124-7B2026.0617