s5_e1
done · peak unseen_mismatch 93.3
α schedule
α 5.0 ×50 α 4.8261 ×50 α 4.6522 ×50 α 4.4783 ×50 α 4.3043 ×50 α 4.1304 ×50 α 3.9565 ×50 α 3.7826 ×50 α 3.6087 ×50 α 3.4348 ×50 α 3.2609 ×50 α 3.087 ×50 α 2.913 ×50 α 2.7391 ×50 α 2.5652 ×50 α 2.3913 ×50 α 2.2174 ×50 α 2.0435 ×50 α 1.8696 ×50 α 1.6957 ×50 α 1.5217 ×50 α 1.3478 ×50 α 1.1739 ×50 α 1.0 ×50
Figures
Results
accuracy by training iteration, mean ± sd over seeds · schedule ends at α = 1.0
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 | 51.8 ± 2.1 | 49.9 ± 1.2 | 50.2 ± 1.8 | 50.0 ± 1.4 |
| 100 | 60.2 ± 7.6 | 58.5 ± 7.1 | 54.0 ± 4.9 | 52.7 ± 4.2 |
| 150 | 76.0 ± 6.1 | 74.2 ± 6.0 | 62.1 ± 4.5 | 59.5 ± 4.3 |
| 200 | 80.3 ± 1.8 | 79.0 ± 1.3 | 63.8 ± 3.8 | 61.8 ± 3.0 |
| 250 | 83.0 ± 2.1 | 81.8 ± 1.8 | 64.2 ± 3.5 | 62.4 ± 2.9 |
| 300 | 84.7 ± 1.8 | 83.4 ± 1.9 | 66.0 ± 2.8 | 64.5 ± 2.7 |
| 350 | 88.7 ± 2.1 | 86.2 ± 3.0 | 71.2 ± 4.9 | 68.0 ± 5.4 |
| 400 | 87.1 ± 2.1 | 85.1 ± 2.1 | 68.5 ± 4.3 | 63.8 ± 5.5 |
| 450 | 89.4 ± 1.2 | 87.0 ± 1.6 | 71.4 ± 3.9 | 67.3 ± 5.8 |
| 500 | 90.5 ± 1.5 | 88.2 ± 1.8 | 72.5 ± 4.7 | 67.4 ± 5.6 |
| 550 | 90.7 ± 1.7 | 88.5 ± 2.0 | 73.2 ± 5.4 | 69.4 ± 5.1 |
| 600 | 92.0 ± 1.2 | 89.7 ± 1.3 | 75.0 ± 3.2 | 69.8 ± 4.7 |
| 650 | 93.4 ± 1.7 | 90.8 ± 1.6 | 78.5 ± 5.4 | 71.4 ± 6.2 |
| 700 | 93.4 ± 1.4 | 91.1 ± 2.0 | 77.7 ± 4.3 | 70.9 ± 4.6 |
| 750 | 94.7 ± 0.9 | 92.5 ± 1.7 | 81.6 ± 5.5 | 74.5 ± 7.9 |
| 800 | 95.6 ± 1.3 | 93.4 ± 1.7 | 84.3 ± 4.9 | 75.9 ± 6.1 |
| 850 | 96.8 ± 0.8 | 94.3 ± 1.5 | 87.7 ± 4.3 | 77.6 ± 5.2 |
| 900 | 97.9 ± 0.6 | 96.0 ± 1.4 | 87.8 ± 3.2 | 80.5 ± 5.3 |
| 950 | 98.1 ± 0.7 | 96.3 ± 1.4 | 91.4 ± 3.4 | 81.7 ± 5.2 |
| 1000 | 99.1 ± 0.4 | 97.1 ± 1.2 | 94.0 ± 1.8 | 82.4 ± 6.1 |
| 1050 | 99.5 ± 0.5 | 98.3 ± 0.9 | 95.8 ± 2.1 | 85.9 ± 5.9 |
| 1100 | 99.7 ± 0.2 | 98.2 ± 1.2 | 97.4 ± 0.8 | 86.8 ± 6.0 |
| 1150 | 100.0 ± 0.1 | 99.2 ± 0.8 | 99.5 ± 0.4 | 93.3 ± 5.1 |
| 1200 | 100.0 ± 0.0 | 99.0 ± 1.5 | 99.5 ± 0.4 | 91.2 ± 4.7 |