List your company

Instruction-following on Dolly (Diversity and Pass@K Metrics)

36Distinct-1GRPO-entropy
23.5226.763033.24Sep 5, 2025

Evaluation results

MethodLinks
GRPO-entropyobjective=entropy2025.09367556574.717.7
DQO2025.09286946545.928.74
GRPO-likelihoodobjective=likelihood2025.09267046545.868.5
PPO2025.09246441495.658.39