Exact Retrieval on Needle-in-a-Haystack (NIAH) 32K
100Average AccuracyBaseline SFT
Evaluation Results
| Method | Links | |
|---|---|---|
| Baseline SFT2026.04 | 100 | |
| MEAN0.1Anchor=MEAN, p=0.12026.04 | 100 | |
| MIN0.1Anchor=MIN, p=0.12026.04 | 99.93 | |
| MIN0.2Anchor=MIN, p=0.22026.04 | 99.93 | |
| MEAN0.2Anchor=MEAN, p=0.22026.04 | 99.93 |