s1_e0
done · peak unseen_mismatch 50.2
α schedule
α 1.0 ×50 α 0.9565 ×50 α 0.913 ×50 α 0.8696 ×50 α 0.8261 ×50 α 0.7826 ×50 α 0.7391 ×50 α 0.6957 ×50 α 0.6522 ×50 α 0.6087 ×50 α 0.5652 ×50 α 0.5217 ×50 α 0.4783 ×50 α 0.4348 ×50 α 0.3913 ×50 α 0.3478 ×50 α 0.3043 ×50 α 0.2609 ×50 α 0.2174 ×50 α 0.1739 ×50 α 0.1304 ×50 α 0.087 ×50 α 0.0435 ×50 α 0.0 ×50
Figures
Results
accuracy by training iteration, mean ± sd over seeds · schedule ends at α = 0.0
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 |
|---|---|---|---|---|
| 50 | 50.6 ± 1.3 | 49.1 ± 0.7 | 50.0 ± 2.0 | 50.2 ± 1.2 |
| 100 | 50.2 ± 1.4 | 49.7 ± 1.9 | 49.6 ± 0.7 | 49.5 ± 1.1 |
| 150 | 50.4 ± 3.0 | 49.0 ± 2.9 | 51.3 ± 3.6 | 49.2 ± 2.3 |
| 200 | 62.6 ± 19.0 | 46.8 ± 7.1 | 60.8 ± 19.6 | 46.8 ± 6.0 |
| 250 | 81.3 ± 19.5 | 37.0 ± 11.6 | 80.4 ± 20.7 | 38.5 ± 11.4 |
| 300 | 90.6 ± 18.9 | 34.8 ± 18.6 | 90.5 ± 18.9 | 36.2 ± 18.5 |
| 350 | 96.0 ± 12.0 | 24.0 ± 15.7 | 95.8 ± 12.5 | 23.8 ± 15.9 |
| 400 | 100.0 ± 0.0 | 22.4 ± 19.8 | 100.0 ± 0.0 | 23.3 ± 20.1 |
| 450 | 100.0 ± 0.0 | 31.9 ± 20.5 | 100.0 ± 0.0 | 32.9 ± 20.9 |
| 500 | 100.0 ± 0.0 | 27.6 ± 20.9 | 100.0 ± 0.0 | 28.6 ± 21.1 |
| 550 | 100.0 ± 0.0 | 26.2 ± 19.3 | 100.0 ± 0.0 | 26.9 ± 19.5 |
| 600 | 100.0 ± 0.0 | 25.9 ± 15.2 | 100.0 ± 0.0 | 26.2 ± 15.9 |
| 650 | 100.0 ± 0.0 | 38.6 ± 16.1 | 100.0 ± 0.0 | 39.8 ± 16.0 |
| 700 | 100.0 ± 0.0 | 37.1 ± 25.3 | 100.0 ± 0.0 | 37.1 ± 24.4 |
| 750 | 100.0 ± 0.0 | 45.6 ± 29.8 | 100.0 ± 0.0 | 45.6 ± 29.7 |
| 800 | 100.0 ± 0.0 | 39.5 ± 28.5 | 100.0 ± 0.0 | 39.5 ± 28.7 |
| 850 | 100.0 ± 0.0 | 33.4 ± 28.7 | 100.0 ± 0.0 | 33.5 ± 28.4 |
| 900 | 100.0 ± 0.0 | 39.4 ± 24.7 | 100.0 ± 0.0 | 39.6 ± 24.3 |
| 950 | 100.0 ± 0.0 | 27.5 ± 29.5 | 100.0 ± 0.0 | 27.9 ± 29.5 |
| 1000 | 100.0 ± 0.0 | 28.8 ± 18.0 | 100.0 ± 0.0 | 28.9 ± 17.8 |
| 1050 | 100.0 ± 0.0 | 31.0 ± 18.4 | 100.0 ± 0.0 | 31.2 ± 19.2 |
| 1100 | 100.0 ± 0.0 | 26.7 ± 28.3 | 100.0 ± 0.0 | 27.1 ± 28.5 |
| 1150 | 100.0 ± 0.0 | 17.3 ± 29.0 | 100.0 ± 0.0 | 17.1 ± 29.2 |
| 1200 | 100.0 ± 0.0 | 14.8 ± 18.0 | 100.0 ± 0.0 | 14.4 ± 17.9 |