Code Generation on HumanEval 29 tasks
96.72Mean F1 Scorezebra-mult_offset
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| zebra-mult_offsetalpha=0.5, runs_per_cell=152026.05 | 96.72 | 99.1 | 11 | 0.0253 | |
| zebra-additivealpha=0.5, runs_per_cell=152026.05 | 96.26 | 98.6 | 9 | 0.0754 | |
| llm_cotalpha=0.5, runs_per_cell=152026.05 | 94.63 | 97 | 8 | 0.8752 | |
| llmalpha=0.5, runs_per_cell=152026.05 | 94.19 | 96.5 | — | — |