List your company

Open-ended generation on CARE-pro

19.75Score (Seen)GRPO
2.23646.783211.3315.8768May 25, 2026

Evaluation results

MethodLinks
GRPOBackbone=Gemma3-4b-it2026.0519.7517.2518.92
LRPOBackbone=Gemma3-4b-it2026.0519.2616.8418.45
DPOBackbone=Gemma3-4b-it2026.0518.6312.7316.66
MAPOBackbone=Gemma3-4b-it2026.0517.6911.515.62
VanillaBackbone=Gemma3-4b-it2026.0516.7412.9415.47
LIDRBackbone=Gemma3-4b-it2026.0516.2110.0614.16
MPOBackbone=Gemma3-4b-it2026.0515.4110.4713.76
LRPOBackbone=Qwen2.5-1.5b-it2026.0512.361.238.65
DPOBackbone=Qwen2.5-1.5b-it2026.059.61.646.94
VanillaBackbone=Qwen2.5-1.5b-it2026.058.580.825.99
GRPOBackbone=Qwen2.5-1.5b-it2026.058.260.825.78
MPOBackbone=Qwen2.5-1.5b-it2026.056.581.034.73
LIDRBackbone=LLaMA3.2-1b-it2026.056.481.034.66
LIDRBackbone=Qwen2.5-1.5b-it2026.056.440.624.5
MAPOBackbone=Qwen2.5-1.5b-it2026.056.20.624.34
MPOBackbone=LLaMA3.2-1b-it2026.055.533.494.85
LRPOBackbone=LLaMA3.2-1b-it2026.053.963.083.66
VanillaBackbone=LLaMA3.2-1b-it2026.053.52.263.09
MAPOBackbone=LLaMA3.2-1b-it2026.053.52.463.15
DPOBackbone=LLaMA3.2-1b-it2026.053.011.642.55
GRPOBackbone=LLaMA3.2-1b-it2026.052.913.333.05