free-6217439f
done · peak unseen_mismatch 92.6
α schedule
α 5.0 ×10 α 4.1 ×10 α 4.3 ×10 α 4.9 ×10 α 4.5 ×10 α 4.4 ×10 α 4.0 ×10 α 4.5 ×10 α 4.7 ×10 α 5.0 ×20 α 4.3 ×10 α 4.6 ×10 α 4.3 ×10 α 4.8 ×10 α 4.4 ×10 α 4.2 ×10 α 1.2 ×10 α 0.0 ×10 α 0.6 ×10 α 0.0 ×10 α 0.6 ×10 α 2.1 ×10 α 3.8 ×20 α 5.0 ×10 α 1.7 ×10 α 1.3 ×10 α 0.0 ×120
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 |
|---|---|---|---|---|
| 100 | 60.5 ± 7.8 | 56.7 ± 6.2 | 54.2 ± 5.9 | 51.8 ± 3.6 |
| 110 | 65.4 ± 9.2 | 61.6 ± 7.9 | 56.9 ± 7.5 | 53.6 ± 5.3 |
| 120 | 70.0 ± 10.2 | 66.4 ± 9.1 | 59.7 ± 8.8 | 56.2 ± 6.8 |
| 130 | 74.1 ± 9.6 | 70.6 ± 8.6 | 62.0 ± 9.0 | 58.4 ± 7.1 |
| 140 | 76.9 ± 7.9 | 73.7 ± 7.4 | 63.4 ± 8.9 | 60.1 ± 6.9 |
| 150 | 78.6 ± 6.6 | 76.1 ± 5.9 | 64.1 ± 8.2 | 61.1 ± 6.7 |
| 160 | 80.0 ± 5.6 | 77.5 ± 5.0 | 64.4 ± 7.7 | 61.5 ± 6.4 |
| 170 | 80.6 ± 4.7 | 78.3 ± 4.3 | 64.6 ± 7.0 | 62.1 ± 6.0 |
| 180 | 84.6 ± 5.6 | 82.3 ± 5.3 | 70.5 ± 8.8 | 67.4 ± 8.0 |
| 190 | 90.3 ± 7.2 | 87.8 ± 7.5 | 82.4 ± 10.8 | 78.7 ± 10.7 |
| 200 | 95.9 ± 6.4 | 93.8 ± 7.2 | 92.9 ± 8.9 | 89.8 ± 9.5 |
| 210 | 97.3 ± 6.0 | 93.6 ± 9.5 | 95.9 ± 7.5 | 91.0 ± 10.8 |
| 220 | 97.0 ± 4.9 | 83.0 ± 17.7 | 96.1 ± 6.3 | 80.1 ± 17.8 |
| 230 | 98.7 ± 2.7 | 79.4 ± 21.4 | 97.9 ± 3.3 | 76.7 ± 21.3 |
| 240 | 98.7 ± 1.9 | 87.7 ± 12.7 | 97.8 ± 2.4 | 84.0 ± 12.2 |
| 250 | 96.5 ± 2.1 | 93.4 ± 1.9 | 92.5 ± 4.3 | 87.3 ± 3.3 |
| 260 | 93.1 ± 1.3 | 90.5 ± 0.3 | 88.0 ± 3.7 | 81.5 ± 2.7 |
| 270 | 96.4 ± 0.8 | 92.6 ± 1.0 | 92.7 ± 3.7 | 87.4 ± 3.3 |
| 280 | 98.6 ± 0.1 | 95.0 ± 0.6 | 95.5 ± 0.7 | 88.5 ± 1.1 |
| 290 | 99.2 ± 0.2 | 95.0 ± 2.1 | 98.4 ± 1.0 | 90.3 ± 1.6 |
| 300 | 100.0 ± 0.0 | 88.7 ± 10.6 | 100.0 ± 0.0 | 85.3 ± 13.6 |
| 310 | 100.0 ± 0.0 | 95.4 ± 0.0 | 100.0 ± 0.0 | 92.6 ± 0.0 |