List your company

Language Generation on Goodreads Books (test)

61.38ROUGE-1P-GRPO
39.22844.97950.7356.481Feb 17, 2026

Evaluation results

MethodLinks
P-GRPOModel=Qwen3-8B2026.0261.3813.8335.7852.96
P-GRPOBackbone=Qwen3-8B2026.0261.3813.8335.8752.96
GRPOModel=Qwen3-8B2026.0260.7613.7435.9652.66
GRPOBackbone=Qwen3-8B2026.0260.7613.7435.9652.66
P-GRPOModel=Gemma-2B2026.0255.3411.8132.0454.15
P-GRPOBackbone=Gemma-2B2026.0255.3411.8132.0454.15
GRPOModel=Gemma-2B2026.0255.2610.632.5553.28
GRPOBackbone=Gemma-2B2026.0255.2610.632.5553.28
GDPOModel=Qwen3-8B2026.0249.877.8130.0345.2
GDPOBackbone=Qwen3-8B2026.0249.877.8130.0345.2
GDPOModel=Gemma-2B2026.0240.08622.5452.88
GDPOBackbone=Gemma-2B2026.0240.08622.5452.88