r3_c2
done · peak unseen_mismatch 97.8
α schedule
α 3.2723 ×25 α 3.272 ×25 α 3.2701 ×25 α 3.2643 ×25 α 3.251 ×25 α 3.2257 ×25 α 3.1831 ×25 α 3.1166 ×25 α 3.0189 ×25 α 2.8816 ×25 α 2.6956 ×25 α 2.4507 ×25 α 2.1358 ×25 α 1.739 ×25 α 1.2475 ×25 α 0.6476 ×25
Figures
Results
accuracy by training iteration, mean ± sd over seeds · schedule ends at α = 0.6476
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 |
|---|---|---|---|---|
| 25 | 50.8 ± 1.4 | 50.4 ± 1.8 | 49.8 ± 1.7 | 49.5 ± 1.8 |
| 50 | 51.6 ± 1.6 | 50.4 ± 1.6 | 50.1 ± 1.2 | 49.8 ± 1.8 |
| 75 | 52.2 ± 1.7 | 50.2 ± 1.5 | 50.5 ± 1.5 | 49.4 ± 1.7 |
| 100 | 55.8 ± 6.7 | 53.5 ± 3.6 | 52.2 ± 4.4 | 50.3 ± 2.1 |
| 125 | 66.6 ± 12.8 | 62.8 ± 10.9 | 58.3 ± 11.2 | 55.7 ± 9.0 |
| 150 | 76.3 ± 11.2 | 73.2 ± 11.0 | 65.5 ± 11.2 | 62.5 ± 10.3 |
| 175 | 84.1 ± 7.1 | 81.4 ± 6.9 | 71.2 ± 10.4 | 67.8 ± 9.8 |
| 200 | 87.1 ± 4.3 | 84.6 ± 4.6 | 72.8 ± 9.1 | 70.2 ± 8.8 |
| 225 | 88.5 ± 3.5 | 86.2 ± 3.4 | 73.6 ± 7.2 | 70.9 ± 7.1 |
| 250 | 89.2 ± 2.1 | 87.1 ± 2.0 | 73.6 ± 5.0 | 71.3 ± 5.4 |
| 275 | 90.8 ± 2.4 | 88.9 ± 2.2 | 76.3 ± 4.9 | 73.7 ± 5.1 |
| 300 | 92.6 ± 2.3 | 90.8 ± 2.1 | 79.0 ± 5.1 | 75.7 ± 4.9 |
| 325 | 94.8 ± 1.4 | 92.8 ± 1.6 | 82.1 ± 4.7 | 78.9 ± 5.1 |
| 350 | 97.2 ± 1.1 | 96.0 ± 1.4 | 89.0 ± 3.4 | 85.2 ± 3.8 |
| 375 | 99.3 ± 0.5 | 98.7 ± 0.6 | 96.0 ± 2.0 | 93.5 ± 2.7 |
| 400 | 100.0 ± 0.0 | 99.3 ± 2.4 | 100.0 ± 0.1 | 97.8 ± 4.7 |