free-74bd9917
done · peak unseen_mismatch 82.3
α schedule
α 3.344 ×31 α 1.656 ×16 α 3.38 ×19 α 2.01 ×28 α 3.994 ×11 α 1.519 ×23 α 3.674 ×12 α 1.534 ×16 α 3.657 ×17 α 1.472 ×34 α 3.669 ×15 α 1.964 ×29 α 3.96 ×49 α 3.96 ×100
Results
accuracy by training iteration, mean ± sd over seeds · schedule ends at α = 3.96
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 |
|---|---|---|---|---|
| 100 | 54.1 ± 4.6 | 50.4 ± 2.2 | 52.2 ± 4.0 | 49.5 ± 1.8 |
| 110 | 56.7 ± 6.8 | 51.5 ± 3.1 | 53.9 ± 6.7 | 49.9 ± 2.5 |
| 120 | 59.5 ± 10.4 | 52.6 ± 5.1 | 56.1 ± 10.2 | 50.6 ± 4.1 |
| 130 | 64.3 ± 13.4 | 54.7 ± 6.7 | 60.5 ± 13.4 | 52.1 ± 5.6 |
| 140 | 70.9 ± 16.2 | 59.6 ± 9.8 | 66.5 ± 16.3 | 55.8 ± 8.7 |
| 150 | 75.8 ± 17.4 | 63.9 ± 11.2 | 72.2 ± 17.9 | 59.7 ± 9.9 |
| 160 | 81.9 ± 16.8 | 69.0 ± 13.6 | 78.6 ± 18.1 | 65.0 ± 13.1 |
| 170 | 85.9 ± 15.9 | 75.8 ± 14.8 | 82.2 ± 17.3 | 71.2 ± 14.5 |
| 180 | 88.2 ± 15.0 | 80.3 ± 14.9 | 84.5 ± 16.2 | 75.7 ± 14.7 |
| 190 | 90.7 ± 13.9 | 82.3 ± 14.7 | 87.7 ± 15.6 | 78.1 ± 14.3 |
| 200 | 91.2 ± 13.5 | 81.6 ± 14.1 | 88.8 ± 15.1 | 77.9 ± 13.7 |
| 210 | 92.8 ± 12.3 | 80.9 ± 15.3 | 90.3 ± 13.9 | 77.6 ± 14.7 |
| 220 | 93.3 ± 11.5 | 83.3 ± 13.4 | 89.3 ± 13.2 | 78.8 ± 13.1 |
| 230 | 93.3 ± 10.4 | 87.9 ± 11.1 | 89.2 ± 13.0 | 82.3 ± 12.7 |
| 240 | 93.0 ± 11.3 | 86.9 ± 13.4 | 89.3 ± 12.1 | 82.0 ± 12.3 |
| 250 | 93.7 ± 8.7 | 86.0 ± 12.0 | 88.9 ± 10.6 | 79.7 ± 10.1 |
| 260 | 94.0 ± 6.2 | 90.0 ± 7.1 | 88.0 ± 8.0 | 82.3 ± 7.5 |
| 270 | 92.5 ± 4.0 | 89.2 ± 4.4 | 84.4 ± 7.1 | 78.6 ± 6.9 |
| 280 | 92.4 ± 3.4 | 89.3 ± 3.9 | 84.3 ± 6.2 | 78.7 ± 7.7 |
| 290 | 92.1 ± 2.2 | 89.6 ± 2.3 | 83.2 ± 4.9 | 77.4 ± 6.7 |
| 300 | 91.7 ± 2.3 | 88.9 ± 1.5 | 82.0 ± 5.1 | 76.6 ± 6.8 |
| 310 | 91.5 ± 2.0 | 88.8 ± 2.4 | 81.5 ± 5.4 | 76.2 ± 7.3 |
| 320 | 91.4 ± 2.2 | 88.7 ± 2.5 | 81.4 ± 5.7 | 77.0 ± 6.9 |
| 330 | 91.5 ± 2.1 | 88.9 ± 2.7 | 80.6 ± 5.4 | 75.6 ± 7.5 |
| 340 | 90.3 ± 1.2 | 88.1 ± 2.6 | 78.6 ± 5.0 | 73.7 ± 7.5 |
| 350 | 90.7 ± 2.4 | 88.4 ± 2.6 | 79.7 ± 5.9 | 75.3 ± 7.1 |
| 360 | 91.0 ± 2.2 | 88.4 ± 2.9 | 78.5 ± 5.3 | 74.1 ± 7.4 |
| 370 | 90.5 ± 2.1 | 88.4 ± 3.0 | 78.7 ± 6.1 | 73.3 ± 7.5 |
| 380 | 90.1 ± 2.0 | 87.8 ± 3.0 | 78.0 ± 5.6 | 73.0 ± 7.1 |
| 390 | 90.5 ± 1.5 | 87.9 ± 2.7 | 78.5 ± 4.6 | 73.1 ± 7.1 |
| 400 | 90.6 ± 1.8 | 88.2 ± 3.1 | 78.3 ± 4.7 | 72.6 ± 7.8 |