Code Generation on LeetCodeDataset (all tasks)
25.44Pass@1ACES-C
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| ACES-CBackbone=Qwen2.5-Coder-14B2026.04 | 25.44 | 26.75 | 30.7 | |
| ACES-OBackbone=Qwen2.5-Coder-14B2026.04 | 22.37 | 25 | 28.95 | |
| ACES-CBackbone=DeepSeek-Coder-V2-16B2026.04 | 21.15 | 22.47 | 24.23 | |
| Zero-shotBackbone=DeepSeek-Coder-V2-16B2026.04 | 20.68 | 24.25 | 28.14 | |
| Majority VotingBackbone=Qwen2.5-Coder-14B2026.04 | 20.61 | 24.12 | 29.39 | |
| Zero-shotBackbone=Qwen2.5-Coder-14B2026.04 | 20.54 | 24.94 | 30.48 | |
| ACES-OBackbone=DeepSeek-Coder-V2-16B2026.04 | 18.5 | 20.26 | 23.79 | |
| Majority VotingBackbone=DeepSeek-Coder-V2-16B2026.04 | 16.74 | 18.5 | 22.91 | |
| ACES-CBackbone=Qwen2.5-Coder-7B2026.04 | 16.23 | 17.54 | 18.86 | |
| Zero-shotBackbone=Qwen2.5-Coder-7B2026.04 | 14.64 | 18.82 | 24.5 | |
| ACES-OBackbone=Qwen2.5-Coder-7B2026.04 | 14.04 | 16.23 | 18.86 | |
| Majority VotingBackbone=Qwen2.5-Coder-7B2026.04 | 13.16 | 14.91 | 17.98 |