free-e059ed86
done · peak unseen_mismatch 94.8
α schedule
α 4.8 ×10 α 4.3 ×10 α 5.0 ×10 α 4.0 ×10 α 4.3 ×10 α 4.2 ×10 α 4.8 ×10 α 4.0 ×10 α 4.6 ×10 α 4.0 ×10 α 4.7 ×10 α 4.4 ×10 α 4.1 ×10 α 4.8 ×10 α 5.0 ×10 α 4.5 ×10 α 4.4 ×10 α 1.1 ×10 α 0.8 ×10 α 0.3 ×10 α 0.4 ×10 α 0.9 ×10 α 1.9 ×10 α 5.0 ×10 α 4.2 ×10 α 4.5 ×10 α 2.1 ×10 α 1.4 ×10 α 0.0 ×10 α 0.4 ×110
Figures
Results
accuracy by training iteration, mean ± sd over seeds · schedule ends at α = 0.4
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 | 59.9 ± 7.5 | 56.5 ± 6.1 | 54.1 ± 5.6 | 51.6 ± 3.4 |
| 110 | 65.0 ± 9.0 | 61.1 ± 7.9 | 56.4 ± 7.6 | 53.6 ± 5.0 |
| 120 | 70.0 ± 9.6 | 66.2 ± 8.5 | 59.5 ± 8.7 | 56.2 ± 6.5 |
| 130 | 74.3 ± 9.5 | 70.8 ± 8.6 | 62.2 ± 9.0 | 59.0 ± 7.0 |
| 140 | 77.3 ± 8.4 | 74.4 ± 7.6 | 64.2 ± 8.9 | 60.8 ± 7.1 |
| 150 | 79.0 ± 6.8 | 76.4 ± 6.0 | 64.6 ± 8.2 | 61.6 ± 6.7 |
| 160 | 80.0 ± 5.5 | 77.6 ± 5.0 | 64.4 ± 7.1 | 61.7 ± 6.1 |
| 170 | 80.8 ± 4.6 | 78.4 ± 4.0 | 64.9 ± 6.8 | 62.2 ± 5.8 |
| 180 | 84.7 ± 5.6 | 82.2 ± 5.2 | 70.9 ± 8.8 | 67.8 ± 7.9 |
| 190 | 91.5 ± 5.9 | 89.4 ± 6.2 | 83.0 ± 9.9 | 79.6 ± 9.9 |
| 200 | 96.3 ± 4.9 | 93.9 ± 6.2 | 93.1 ± 7.1 | 89.7 ± 8.3 |
| 210 | 98.1 ± 3.0 | 92.9 ± 9.4 | 97.0 ± 4.2 | 90.3 ± 10.4 |
| 220 | 99.5 ± 0.7 | 86.5 ± 13.9 | 99.0 ± 1.8 | 83.2 ± 14.4 |
| 230 | 99.8 ± 0.4 | 84.9 ± 12.7 | 99.6 ± 0.4 | 82.1 ± 12.5 |
| 240 | 99.8 ± 0.2 | 86.5 ± 8.9 | 99.5 ± 0.8 | 82.4 ± 9.6 |
| 250 | 97.8 ± 2.6 | 86.7 ± 10.0 | 95.5 ± 5.1 | 79.0 ± 11.7 |
| 260 | 97.5 ± 1.4 | 90.1 ± 5.2 | 93.8 ± 3.1 | 81.7 ± 8.0 |
| 270 | 98.2 ± 1.2 | 92.1 ± 3.5 | 95.3 ± 2.1 | 83.8 ± 8.3 |
| 280 | 99.2 ± 0.4 | 95.7 ± 2.4 | 98.5 ± 1.0 | 91.9 ± 4.4 |
| 290 | 99.9 ± 0.1 | 96.7 ± 1.9 | 100.0 ± 0.0 | 94.8 ± 2.8 |