r7_c1
done · peak unseen_mismatch 94.1
α schedule
α 5.0 ×25 α 5.0 ×25 α 5.0 ×25 α 0.6063 ×25 α 0.5661 ×25 α 5.0 ×25 α 5.0 ×25 α 4.5467 ×25 α 0.365 ×25 α 0.3063 ×25 α 0.2446 ×25 α 1.9461 ×25 α 1.2196 ×25 α 0.0431 ×25
Figures
Results
accuracy by training iteration, mean ± sd over seeds · schedule ends at α = 0.0431
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 | 50.9 ± 1.3 | 50.5 ± 1.6 | 49.9 ± 1.4 | 49.7 ± 2.0 |
| 50 | 51.5 ± 2.0 | 50.5 ± 1.5 | 50.0 ± 1.3 | 50.0 ± 1.6 |
| 75 | 53.4 ± 2.6 | 50.7 ± 1.7 | 50.4 ± 1.3 | 49.6 ± 1.5 |
| 100 | 54.9 ± 8.2 | 51.6 ± 2.8 | 52.8 ± 6.7 | 50.1 ± 2.3 |
| 125 | 61.7 ± 15.1 | 50.1 ± 2.7 | 59.4 ± 15.7 | 48.7 ± 3.4 |
| 150 | 75.7 ± 18.9 | 64.7 ± 13.6 | 73.4 ± 19.7 | 61.5 ± 12.5 |
| 175 | 85.8 ± 13.2 | 79.7 ± 12.7 | 81.6 ± 14.5 | 75.2 ± 13.3 |
| 200 | 91.4 ± 6.1 | 87.6 ± 7.2 | 85.2 ± 9.7 | 80.5 ± 9.7 |
| 225 | 99.6 ± 0.7 | 92.2 ± 11.5 | 99.4 ± 1.0 | 91.2 ± 12.1 |
| 250 | 100.0 ± 0.0 | 88.8 ± 16.7 | 100.0 ± 0.1 | 88.1 ± 16.9 |
| 275 | 100.0 ± 0.0 | 80.0 ± 22.7 | 100.0 ± 0.0 | 79.6 ± 22.8 |
| 300 | 100.0 ± 0.0 | 92.6 ± 14.2 | 100.0 ± 0.0 | 92.5 ± 14.0 |
| 325 | 100.0 ± 0.1 | 94.3 ± 13.2 | 99.9 ± 0.3 | 94.1 ± 12.5 |
| 350 | 100.0 ± 0.0 | 93.0 ± 12.4 | 100.0 ± 0.0 | 92.9 ± 12.4 |