Code Generation on APPS+ Interview
12.14Pass@1WITH GRPO ADVANTAGE
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| WITH GRPO ADVANTAGEBackbone=Qwen-2.5-Coder, Parameters=7B2026.05 | 12.14 | 30.2 | |
| SFTBackbone=Qwen-2.5-Coder, Parameters=7B2026.05 | 11.3 | 30.3 | |
| ADVATNAGE WITH EXP()Backbone=Qwen-2.5-Coder, Parameters=7B2026.05 | 11.09 | 30.3 | |
| RLOOBackbone=Qwen-2.5-Coder, Parameters=7B2026.05 | 10.31 | 27.8 | |
| BASEBackbone=Qwen-2.5-Coder, Parameters=7B2026.05 | 7.01 | 25.7 |