free-95c14102
done · peak unseen_mismatch 97.2
α schedule
α 4.164 ×170 α 0.0 ×32 α 4.458 ×80 α 0.7149 ×18 α 0.7149 ×100
Results
accuracy by training iteration, mean ± sd over seeds · schedule ends at α = 0.7149
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 | 61.2 ± 8.4 | 57.0 ± 6.5 | 54.4 ± 6.4 | 51.8 ± 3.6 |
| 110 | 65.8 ± 9.9 | 61.9 ± 8.5 | 56.5 ± 8.4 | 53.7 ± 6.0 |
| 120 | 70.6 ± 10.7 | 67.0 ± 9.3 | 59.7 ± 9.3 | 56.6 ± 7.3 |
| 130 | 74.7 ± 9.7 | 71.1 ± 9.3 | 62.1 ± 9.3 | 58.7 ± 8.1 |
| 140 | 78.2 ± 7.9 | 74.3 ± 7.7 | 64.3 ± 8.8 | 60.6 ± 7.3 |
| 150 | 79.9 ± 6.5 | 76.6 ± 5.9 | 64.7 ± 8.5 | 61.5 ± 7.1 |
| 160 | 81.1 ± 5.3 | 78.0 ± 4.9 | 65.2 ± 8.7 | 62.4 ± 6.9 |
| 170 | 81.6 ± 5.2 | 78.8 ± 4.8 | 65.8 ± 8.1 | 63.0 ± 6.6 |
| 180 | 84.6 ± 7.2 | 81.8 ± 7.1 | 72.6 ± 10.3 | 69.3 ± 9.2 |
| 190 | 91.2 ± 8.7 | 87.9 ± 9.0 | 85.5 ± 11.2 | 81.0 ± 10.7 |
| 200 | 96.7 ± 7.7 | 94.2 ± 8.1 | 95.3 ± 8.8 | 91.6 ± 9.1 |
| 210 | 96.8 ± 9.4 | 93.5 ± 10.1 | 95.7 ± 9.8 | 91.1 ± 10.6 |
| 220 | 91.9 ± 10.4 | 87.5 ± 12.3 | 86.8 ± 12.2 | 81.1 ± 13.4 |
| 230 | 86.4 ± 3.5 | 80.3 ± 4.1 | 71.9 ± 6.7 | 66.8 ± 2.2 |
| 240 | 84.6 ± 2.5 | 78.7 ± 3.4 | 69.2 ± 6.3 | 65.7 ± 2.1 |
| 250 | 87.8 ± 4.0 | 83.1 ± 1.8 | 73.0 ± 8.9 | 71.0 ± 4.8 |
| 260 | 86.9 ± 2.2 | 83.4 ± 0.8 | 73.2 ± 7.2 | 71.2 ± 4.5 |
| 270 | 87.5 ± 2.2 | 83.9 ± 2.0 | 73.0 ± 7.1 | 70.6 ± 5.8 |
| 280 | 88.2 ± 3.3 | 86.3 ± 3.9 | 73.8 ± 8.7 | 72.3 ± 8.1 |
| 290 | 91.9 ± 3.0 | 89.3 ± 3.1 | 80.9 ± 8.3 | 78.7 ± 7.7 |
| 300 | 97.3 ± 1.9 | 95.4 ± 1.8 | 93.4 ± 5.1 | 90.5 ± 4.1 |
| 310 | 99.8 ± 0.1 | 99.3 ± 0.1 | 98.3 ± 1.2 | 97.2 ± 0.7 |