smooth_16
done · peak unseen_mismatch 98.1
α schedule
α 3.2007 ×25 α 3.1661 ×25 α 3.1038 ×25 α 3.016 ×25 α 2.904 ×25 α 2.7689 ×25 α 2.6113 ×25 α 2.4318 ×25 α 2.2309 ×25 α 2.009 ×25 α 1.7666 ×25 α 1.5038 ×25 α 1.221 ×25 α 0.9185 ×25 α 0.5965 ×25 α 0.2551 ×25
Figures
Results
accuracy by training iteration, mean ± sd over seeds · schedule ends at α = 0.2551
seen_* probed at reference α = 1.4 · unseen_* = held-out pairings, uniform (α-independent) · 1000 pairs per condition
| iteration | seen_match | seen_mismatch | unseen_match | unseen_mismatch |
|---|---|---|---|---|
| 25 | 51.3 ± 1.2 | 50.5 ± 1.2 | 49.5 ± 2.0 | 50.3 ± 1.6 |
| 50 | 51.4 ± 1.3 | 50.2 ± 1.3 | 49.9 ± 1.8 | 49.5 ± 1.6 |
| 75 | 52.0 ± 1.8 | 50.0 ± 1.6 | 50.7 ± 2.2 | 49.8 ± 1.9 |
| 100 | 57.5 ± 7.7 | 53.8 ± 4.0 | 53.6 ± 8.0 | 50.5 ± 1.8 |
| 125 | 70.2 ± 11.8 | 64.9 ± 11.2 | 62.0 ± 11.6 | 56.8 ± 8.9 |
| 150 | 82.6 ± 10.4 | 77.8 ± 10.1 | 73.6 ± 13.2 | 67.5 ± 11.2 |
| 175 | 89.0 ± 7.9 | 86.1 ± 7.9 | 79.2 ± 11.9 | 74.4 ± 10.9 |
| 200 | 91.4 ± 5.6 | 89.0 ± 6.7 | 81.2 ± 10.0 | 77.7 ± 9.6 |
| 225 | 93.6 ± 3.5 | 91.8 ± 3.8 | 84.3 ± 8.4 | 80.0 ± 8.8 |
| 250 | 95.7 ± 2.0 | 94.3 ± 2.3 | 87.9 ± 6.0 | 84.1 ± 6.5 |
| 275 | 97.2 ± 1.2 | 95.7 ± 1.5 | 90.8 ± 4.1 | 87.0 ± 4.2 |
| 300 | 98.5 ± 0.8 | 97.6 ± 1.1 | 94.7 ± 2.6 | 90.8 ± 3.3 |
| 325 | 99.6 ± 0.3 | 99.0 ± 0.6 | 97.6 ± 2.0 | 94.4 ± 3.4 |
| 350 | 99.9 ± 0.1 | 99.6 ± 0.3 | 99.3 ± 0.7 | 97.0 ± 2.4 |
| 375 | 100.0 ± 0.0 | 99.7 ± 0.4 | 99.9 ± 0.2 | 98.1 ± 2.4 |
| 400 | 100.0 ± 0.0 | 90.9 ± 17.7 | 100.0 ± 0.0 | 86.4 ± 20.8 |