Science Question Answering on GPQA Diamond (Score)
70ScoreBaseline
Evaluation Results
| Method | Links | |
|---|---|---|
| BaselinePrecision=BF162026.07 | 70 | |
| AFM-k5984d3s (Ours)Optimization=QAD + Speculative Decoding + SWA2026.07 | 66.7 |
| Method | Links | |
|---|---|---|
| BaselinePrecision=BF162026.07 | 70 | |
| AFM-k5984d3s (Ours)Optimization=QAD + Speculative Decoding + SWA2026.07 | 66.7 |