s5_e4_blk
done · peak unseen_mismatch 68.8
α schedule
α 5.0 ×600 α 4.0 ×600
Figures
Results
accuracy by training iteration, mean ± sd over seeds · schedule ends at α = 4.0
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 | 59.6 ± 6.4 | 56.9 ± 4.6 | 52.8 ± 3.7 | 51.7 ± 3.0 |
| 150 | 76.5 ± 5.5 | 74.3 ± 5.6 | 62.2 ± 5.9 | 59.1 ± 4.4 |
| 200 | 79.5 ± 1.9 | 78.0 ± 2.0 | 60.5 ± 3.5 | 59.9 ± 3.0 |
| 250 | 79.8 ± 1.8 | 78.6 ± 1.6 | 60.5 ± 3.1 | 59.4 ± 3.1 |
| 300 | 82.4 ± 1.6 | 81.0 ± 1.8 | 63.4 ± 3.6 | 62.4 ± 3.8 |
| 350 | 83.8 ± 2.3 | 82.1 ± 2.2 | 65.8 ± 5.6 | 65.2 ± 5.9 |
| 400 | 83.3 ± 1.7 | 81.8 ± 1.5 | 64.7 ± 2.3 | 62.0 ± 3.1 |
| 450 | 83.3 ± 1.8 | 81.3 ± 1.9 | 64.3 ± 3.3 | 61.6 ± 2.8 |
| 500 | 85.0 ± 2.0 | 83.3 ± 2.6 | 66.1 ± 5.5 | 63.1 ± 5.4 |
| 550 | 84.0 ± 2.5 | 82.9 ± 2.3 | 64.5 ± 4.6 | 61.7 ± 4.8 |
| 600 | 84.8 ± 2.1 | 83.4 ± 2.4 | 66.0 ± 3.4 | 62.6 ± 4.7 |
| 650 | 86.8 ± 2.3 | 85.1 ± 1.9 | 67.7 ± 5.8 | 64.5 ± 6.0 |
| 700 | 88.1 ± 1.2 | 85.5 ± 1.6 | 70.1 ± 3.5 | 65.7 ± 4.1 |
| 750 | 87.5 ± 1.2 | 86.0 ± 1.3 | 69.3 ± 4.5 | 63.8 ± 4.2 |
| 800 | 88.3 ± 1.8 | 86.8 ± 2.0 | 69.8 ± 4.9 | 66.7 ± 4.8 |
| 850 | 89.2 ± 1.2 | 87.0 ± 1.5 | 70.9 ± 4.2 | 67.4 ± 4.7 |
| 900 | 88.8 ± 1.3 | 87.3 ± 1.3 | 70.2 ± 5.1 | 65.9 ± 4.7 |
| 950 | 88.8 ± 1.7 | 87.2 ± 1.4 | 70.6 ± 4.5 | 66.6 ± 4.3 |
| 1000 | 89.1 ± 1.4 | 87.9 ± 1.3 | 71.0 ± 4.3 | 67.4 ± 4.8 |
| 1050 | 89.5 ± 1.6 | 88.0 ± 1.6 | 71.8 ± 3.4 | 67.5 ± 4.5 |
| 1100 | 89.7 ± 1.5 | 87.7 ± 0.9 | 72.9 ± 3.9 | 67.6 ± 3.8 |
| 1150 | 89.9 ± 1.1 | 88.2 ± 1.7 | 72.0 ± 3.5 | 67.5 ± 4.7 |
| 1200 | 90.5 ± 1.2 | 88.1 ± 1.4 | 73.2 ± 3.0 | 68.8 ± 3.6 |