s5_e2_blk
done · peak unseen_mismatch 82.3
α schedule
α 5.0 ×300 α 4.0 ×300 α 3.0 ×300 α 2.0 ×300
Figures
Results
accuracy by training iteration, mean ± sd over seeds · schedule ends at α = 2.0
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 | 85.8 ± 3.1 | 84.5 ± 2.7 | 67.5 ± 5.8 | 65.3 ± 5.9 |
| 400 | 86.0 ± 1.4 | 83.8 ± 1.9 | 66.4 ± 3.1 | 63.3 ± 4.6 |
| 450 | 86.4 ± 2.0 | 85.3 ± 2.0 | 67.0 ± 5.2 | 64.4 ± 5.3 |
| 500 | 87.3 ± 2.1 | 85.8 ± 2.0 | 67.0 ± 4.6 | 64.0 ± 4.1 |
| 550 | 88.3 ± 1.4 | 85.9 ± 1.9 | 70.4 ± 3.7 | 65.8 ± 5.0 |
| 600 | 88.2 ± 2.1 | 86.0 ± 2.1 | 69.0 ± 4.8 | 64.5 ± 5.2 |
| 650 | 90.9 ± 1.2 | 88.4 ± 1.6 | 73.0 ± 2.9 | 68.6 ± 4.3 |
| 700 | 91.9 ± 1.4 | 89.5 ± 1.8 | 74.3 ± 5.1 | 68.2 ± 6.4 |
| 750 | 92.6 ± 1.3 | 90.8 ± 1.5 | 76.3 ± 4.6 | 70.6 ± 5.3 |
| 800 | 92.6 ± 1.1 | 90.7 ± 1.4 | 77.0 ± 3.8 | 71.3 ± 4.6 |
| 850 | 92.5 ± 1.2 | 90.5 ± 1.3 | 76.5 ± 3.6 | 70.6 ± 4.8 |
| 900 | 93.3 ± 1.2 | 91.4 ± 1.3 | 78.5 ± 4.1 | 72.3 ± 5.0 |
| 950 | 97.0 ± 0.9 | 95.9 ± 1.6 | 86.1 ± 4.2 | 78.8 ± 7.3 |
| 1000 | 98.0 ± 0.9 | 96.4 ± 1.5 | 90.1 ± 3.9 | 82.0 ± 6.8 |
| 1050 | 97.9 ± 0.3 | 96.1 ± 1.4 | 89.5 ± 2.7 | 81.1 ± 6.4 |
| 1100 | 98.4 ± 0.6 | 96.9 ± 0.9 | 90.7 ± 4.1 | 81.9 ± 5.3 |
| 1150 | 98.2 ± 0.7 | 96.3 ± 2.1 | 90.6 ± 3.6 | 81.5 ± 9.0 |
| 1200 | 98.5 ± 0.7 | 96.9 ± 1.3 | 91.0 ± 2.9 | 82.3 ± 7.2 |