free-2d3dce9e
done · peak unseen_mismatch 83.7
α schedule
α 3.194 ×70 α 1.507 ×64 α 3.123 ×12 α 1.255 ×11 α 3.161 ×24 α 1.634 ×24 α 2.729 ×13 α 0.9098 ×12 α 3.345 ×21 α 0.7192 ×22 α 2.79 ×12 α 1.251 ×15 α 1.251 ×100
Results
accuracy by training iteration, mean ± sd over seeds · schedule ends at α = 1.251
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 | 53.9 ± 4.5 | 50.7 ± 2.3 | 51.8 ± 3.8 | 49.6 ± 1.7 |
| 110 | 56.2 ± 8.3 | 51.1 ± 3.5 | 53.7 ± 8.0 | 50.0 ± 2.5 |
| 120 | 59.7 ± 12.3 | 52.1 ± 4.6 | 57.2 ± 12.0 | 50.7 ± 3.9 |
| 130 | 64.6 ± 15.3 | 54.0 ± 7.3 | 61.6 ± 15.2 | 52.2 ± 6.5 |
| 140 | 70.1 ± 17.9 | 57.4 ± 10.5 | 67.4 ± 17.8 | 54.7 ± 9.3 |
| 150 | 75.5 ± 17.5 | 62.7 ± 12.9 | 72.8 ± 17.9 | 59.4 ± 11.8 |
| 160 | 81.0 ± 18.0 | 67.2 ± 15.2 | 78.6 ± 18.6 | 64.6 ± 14.2 |
| 170 | 85.3 ± 16.4 | 73.8 ± 15.0 | 82.7 ± 17.4 | 70.5 ± 14.3 |
| 180 | 88.4 ± 14.0 | 79.7 ± 14.4 | 85.6 ± 15.0 | 75.9 ± 14.3 |
| 190 | 90.6 ± 13.2 | 81.1 ± 14.3 | 88.5 ± 13.6 | 77.2 ± 13.9 |
| 200 | 92.9 ± 12.3 | 84.1 ± 13.8 | 91.0 ± 13.1 | 80.6 ± 13.4 |
| 210 | 94.0 ± 12.5 | 84.1 ± 14.0 | 92.1 ± 13.4 | 80.2 ± 13.2 |
| 220 | 93.5 ± 13.3 | 86.3 ± 13.6 | 91.3 ± 13.9 | 82.3 ± 13.3 |
| 230 | 93.7 ± 13.6 | 83.4 ± 15.0 | 92.7 ± 14.0 | 80.2 ± 14.9 |
| 240 | 93.3 ± 13.7 | 81.4 ± 17.3 | 92.4 ± 14.3 | 79.6 ± 16.8 |
| 250 | 91.9 ± 14.9 | 83.7 ± 13.8 | 90.2 ± 15.3 | 80.6 ± 12.7 |
| 260 | 93.0 ± 14.0 | 80.2 ± 15.1 | 91.1 ± 15.2 | 78.0 ± 13.7 |
| 270 | 93.1 ± 17.1 | 81.8 ± 15.5 | 93.3 ± 15.9 | 79.7 ± 14.1 |
| 280 | 90.3 ± 19.4 | 76.3 ± 17.0 | 90.6 ± 18.8 | 75.7 ± 15.4 |
| 290 | 89.0 ± 19.0 | 79.9 ± 18.3 | 89.0 ± 19.0 | 78.8 ± 16.0 |
| 300 | 88.3 ± 16.5 | 82.1 ± 18.1 | 86.3 ± 19.3 | 79.8 ± 19.1 |
| 310 | 85.2 ± 14.8 | 74.1 ± 16.4 | 82.0 ± 17.9 | 73.2 ± 17.8 |
| 320 | 81.3 ± 0.0 | 66.7 ± 0.0 | 77.4 ± 0.0 | 60.4 ± 0.0 |
| 330 | 97.2 ± 0.0 | 69.4 ± 0.0 | 97.3 ± 0.0 | 64.5 ± 0.0 |
| 340 | 100.0 ± 0.0 | 67.8 ± 0.0 | 99.6 ± 0.0 | 67.0 ± 0.0 |
| 350 | 100.0 ± 0.0 | 83.7 ± 0.0 | 100.0 ± 0.0 | 83.7 ± 0.0 |
| 360 | 100.0 ± 0.0 | 60.9 ± 0.0 | 100.0 ± 0.0 | 60.4 ± 0.0 |
| 370 | 100.0 ± 0.0 | 54.1 ± 0.0 | 100.0 ± 0.0 | 55.4 ± 0.0 |
| 380 | 100.0 ± 0.0 | 75.0 ± 0.0 | 100.0 ± 0.0 | 73.5 ± 0.0 |
| 390 | 100.0 ± 0.0 | 77.8 ± 0.0 | 100.0 ± 0.0 | 77.7 ± 0.0 |
| 400 | 100.0 ± 0.0 | 82.5 ± 0.0 | 99.9 ± 0.0 | 83.6 ± 0.0 |