free-241ff30e
done · peak unseen_mismatch 93.4
α schedule
α 4.9 ×10 α 4.3 ×10 α 5.0 ×10 α 4.7 ×10 α 5.0 ×10 α 4.8 ×10 α 3.6 ×10 α 4.4 ×10 α 5.0 ×10 α 4.3 ×10 α 4.8 ×10 α 4.0 ×10 α 4.7 ×10 α 4.0 ×10 α 5.0 ×20 α 4.2 ×10 α 1.0 ×10 α 0.0 ×10 α 0.4 ×10 α 0.0 ×20 α 2.5 ×10 α 3.3 ×10 α 4.0 ×10 α 5.0 ×10 α 3.0 ×10 α 0.4 ×10 α 0.0 ×10 α 0.1 ×110
Figures
Results
accuracy by training iteration, mean ± sd over seeds · schedule ends at α = 0.1
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 | 61.3 ± 7.8 | 57.6 ± 6.4 | 54.6 ± 5.8 | 51.8 ± 3.5 |
| 110 | 66.5 ± 9.0 | 62.9 ± 7.9 | 57.0 ± 7.2 | 53.9 ± 5.0 |
| 120 | 71.4 ± 9.3 | 67.7 ± 8.7 | 59.9 ± 8.5 | 56.6 ± 6.4 |
| 130 | 75.6 ± 8.5 | 72.1 ± 8.0 | 62.8 ± 8.6 | 59.5 ± 7.0 |
| 140 | 78.1 ± 7.3 | 75.2 ± 6.7 | 64.1 ± 8.3 | 60.7 ± 6.8 |
| 150 | 79.4 ± 6.2 | 76.7 ± 5.7 | 64.3 ± 8.1 | 61.5 ± 6.8 |
| 160 | 79.9 ± 4.9 | 77.5 ± 4.2 | 63.8 ± 7.0 | 61.4 ± 6.1 |
| 170 | 80.6 ± 4.0 | 78.5 ± 3.4 | 64.1 ± 6.4 | 61.8 ± 5.5 |
| 180 | 84.5 ± 5.0 | 82.2 ± 4.6 | 70.1 ± 8.2 | 67.3 ± 7.4 |
| 190 | 90.2 ± 6.6 | 87.7 ± 6.8 | 82.2 ± 10.7 | 79.0 ± 10.4 |
| 200 | 96.5 ± 3.9 | 94.0 ± 5.9 | 93.8 ± 6.2 | 90.3 ± 7.8 |
| 210 | 99.1 ± 1.1 | 94.6 ± 8.9 | 98.4 ± 2.0 | 92.6 ± 9.8 |
| 220 | 99.7 ± 0.7 | 77.5 ± 14.5 | 99.7 ± 0.8 | 75.5 ± 14.2 |
| 230 | 99.9 ± 0.2 | 85.4 ± 14.1 | 99.9 ± 0.3 | 84.0 ± 13.5 |
| 240 | 99.9 ± 0.2 | 89.8 ± 10.6 | 99.7 ± 0.4 | 88.0 ± 10.8 |
| 250 | 99.9 ± 0.1 | 92.5 ± 10.8 | 99.7 ± 0.3 | 90.6 ± 11.2 |
| 260 | 100.0 ± 0.0 | 93.0 ± 0.0 | 99.9 ± 0.0 | 90.0 ± 0.0 |
| 270 | 100.0 ± 0.0 | 97.4 ± 0.0 | 99.8 ± 0.0 | 93.4 ± 0.0 |