peak_30
done · peak unseen_mismatch 97.9
α schedule
α 3.0 ×150 α 1.0 ×150 α 2.1429 ×150 α 1.0 ×150 α 1.2857 ×150 α 1.0 ×150
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 | 50.9 ± 1.9 | 50.4 ± 1.4 | 50.7 ± 1.2 | 50.1 ± 1.3 |
| 100 | 52.8 ± 3.0 | 50.6 ± 1.9 | 51.1 ± 1.7 | 49.4 ± 1.7 |
| 150 | 70.1 ± 16.1 | 62.2 ± 12.3 | 67.1 ± 15.6 | 60.3 ± 10.7 |
| 200 | 94.0 ± 12.4 | 61.9 ± 18.6 | 93.7 ± 13.1 | 60.7 ± 18.0 |
| 250 | 100.0 ± 0.0 | 41.9 ± 19.3 | 100.0 ± 0.0 | 41.9 ± 19.2 |
| 300 | 100.0 ± 0.0 | 40.7 ± 27.0 | 100.0 ± 0.0 | 40.8 ± 27.3 |
| 350 | 100.0 ± 0.0 | 69.2 ± 23.8 | 100.0 ± 0.0 | 69.1 ± 23.7 |
| 400 | 100.0 ± 0.0 | 94.1 ± 15.2 | 100.0 ± 0.1 | 93.7 ± 15.2 |
| 450 | 99.8 ± 0.3 | 93.8 ± 15.1 | 99.5 ± 0.9 | 93.0 ± 14.8 |
| 500 | 100.0 ± 0.0 | 89.5 ± 16.5 | 100.0 ± 0.0 | 89.0 ± 17.4 |
| 550 | 100.0 ± 0.0 | 95.0 ± 13.1 | 100.0 ± 0.1 | 94.1 ± 13.5 |
| 600 | 100.0 ± 0.0 | 97.7 ± 3.1 | 100.0 ± 0.0 | 96.8 ± 3.2 |
| 650 | 100.0 ± 0.0 | 93.2 ± 14.6 | 100.0 ± 0.0 | 91.7 ± 15.2 |
| 700 | 100.0 ± 0.1 | 98.1 ± 2.3 | 99.8 ± 0.5 | 95.8 ± 4.9 |
| 750 | 100.0 ± 0.0 | 97.2 ± 4.1 | 99.7 ± 0.5 | 94.6 ± 5.9 |
| 800 | 100.0 ± 0.0 | 99.5 ± 0.9 | 99.9 ± 0.2 | 97.9 ± 2.1 |
| 850 | 100.0 ± 0.0 | 97.9 ± 2.8 | 99.8 ± 0.6 | 95.8 ± 5.2 |
| 900 | 100.0 ± 0.0 | 98.1 ± 2.5 | 99.7 ± 0.9 | 94.6 ± 5.6 |