s1_e5_blk
done · peak unseen_mismatch 96.3
α schedule
α 1.0 ×240 α 2.0 ×240 α 3.0 ×240 α 4.0 ×240 α 5.0 ×240
Figures
Results
accuracy by training iteration, mean ± sd over seeds · schedule ends at α = 5.0
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 |
|---|---|---|---|---|
| 50 | 50.6 ± 1.3 | 49.1 ± 0.7 | 50.0 ± 2.0 | 50.2 ± 1.2 |
| 100 | 50.1 ± 1.7 | 50.3 ± 2.2 | 50.3 ± 1.8 | 50.0 ± 1.8 |
| 150 | 51.2 ± 1.8 | 49.7 ± 1.9 | 51.1 ± 1.2 | 48.5 ± 1.1 |
| 200 | 57.8 ± 15.6 | 46.8 ± 8.9 | 58.3 ± 15.1 | 47.6 ± 9.7 |
| 250 | 72.6 ± 21.7 | 43.3 ± 8.7 | 72.5 ± 21.3 | 43.4 ± 8.1 |
| 300 | 93.0 ± 13.6 | 44.5 ± 22.8 | 92.3 ± 14.9 | 45.8 ± 22.6 |
| 350 | 99.2 ± 2.3 | 49.0 ± 28.7 | 99.0 ± 2.9 | 50.1 ± 28.0 |
| 400 | 100.0 ± 0.0 | 82.2 ± 15.4 | 100.0 ± 0.0 | 82.0 ± 15.6 |
| 450 | 100.0 ± 0.1 | 79.7 ± 20.9 | 99.9 ± 0.4 | 80.4 ± 19.8 |
| 500 | 99.7 ± 0.6 | 93.3 ± 14.5 | 99.2 ± 1.8 | 92.5 ± 14.2 |
| 550 | 98.9 ± 1.2 | 98.3 ± 1.7 | 97.8 ± 3.3 | 96.3 ± 3.2 |
| 600 | 98.1 ± 1.3 | 96.2 ± 2.4 | 95.8 ± 3.0 | 92.4 ± 3.9 |
| 650 | 96.3 ± 1.6 | 94.0 ± 2.8 | 91.2 ± 4.0 | 87.0 ± 4.2 |
| 700 | 95.6 ± 1.4 | 93.8 ± 2.2 | 89.3 ± 4.0 | 86.0 ± 3.8 |
| 750 | 93.0 ± 1.7 | 91.0 ± 3.3 | 84.4 ± 3.8 | 80.7 ± 4.0 |
| 800 | 91.7 ± 1.6 | 90.0 ± 2.3 | 80.6 ± 3.8 | 77.0 ± 3.4 |
| 850 | 90.7 ± 1.8 | 89.2 ± 2.4 | 78.5 ± 4.1 | 75.5 ± 3.4 |
| 900 | 89.5 ± 1.7 | 87.5 ± 2.3 | 75.3 ± 3.6 | 72.6 ± 3.3 |
| 950 | 88.7 ± 1.6 | 87.2 ± 1.7 | 73.8 ± 3.8 | 71.6 ± 3.3 |
| 1000 | 87.2 ± 1.5 | 85.8 ± 1.5 | 70.8 ± 3.0 | 69.2 ± 1.9 |
| 1050 | 86.9 ± 1.0 | 85.2 ± 1.0 | 69.2 ± 3.2 | 68.1 ± 2.4 |
| 1100 | 86.1 ± 1.3 | 84.7 ± 0.9 | 67.5 ± 3.7 | 67.0 ± 2.8 |
| 1150 | 86.7 ± 1.4 | 84.4 ± 1.7 | 67.9 ± 3.8 | 67.0 ± 3.3 |
| 1200 | 86.4 ± 1.2 | 84.9 ± 1.0 | 67.9 ± 3.6 | 67.1 ± 2.9 |