Hybrid Policy Distillation for LLMs

Open paper
Benchmark Model Rank Results
code-generation-on-mbppHPD (Qwen2.5-Coder, 1.5B)#44Accuracy: 75.4
code-generation-on-humanevalHPD (Qwen2.5-Coder, 1.5B)#63Pass@1: 79.3

Rank counts only results with a code link.