smooth_K48
done · peak unseen_mismatch 91.2
α schedule
α 2.9821 ×19 α 2.9464 ×19 α 2.9107 ×19 α 2.875 ×19 α 2.8393 ×19 α 2.8036 ×19 α 2.7679 ×19 α 2.7321 ×19 α 2.6964 ×19 α 2.6607 ×19 α 2.625 ×19 α 2.5893 ×19 α 2.5536 ×19 α 2.5179 ×19 α 2.4821 ×19 α 2.4464 ×19 α 2.4107 ×19 α 2.375 ×19 α 2.3393 ×19 α 2.3036 ×19 α 2.2679 ×19 α 2.2321 ×19 α 2.1964 ×19 α 2.1607 ×19 α 2.125 ×19 α 2.0893 ×19 α 2.0536 ×19 α 2.0178 ×19 α 1.9821 ×19 α 1.9464 ×19 α 1.9107 ×19 α 1.875 ×19 α 1.8393 ×19 α 1.8036 ×19 α 1.7678 ×19 α 1.7321 ×19 α 1.6964 ×18 α 1.6607 ×18 α 1.625 ×18 α 1.5893 ×18 α 1.5536 ×18 α 1.5178 ×18 α 1.4821 ×18 α 1.4464 ×18 α 1.4107 ×18 α 1.375 ×18 α 1.3393 ×18 α 1.3036 ×18
Figures
Results
accuracy by training iteration, mean ± sd over seeds · schedule ends at α = 1.3036
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 |
|---|---|---|---|---|
| 25 | 50.9 ± 1.7 | 49.9 ± 1.5 | 50.3 ± 1.5 | 50.4 ± 2.4 |
| 50 | 50.9 ± 1.0 | 50.4 ± 1.4 | 49.5 ± 1.3 | 49.7 ± 1.6 |
| 75 | 50.8 ± 1.9 | 49.6 ± 1.2 | 50.5 ± 1.9 | 49.3 ± 1.6 |
| 100 | 54.9 ± 4.0 | 49.5 ± 1.8 | 51.6 ± 2.8 | 50.1 ± 2.0 |
| 125 | 62.0 ± 9.5 | 53.8 ± 4.7 | 58.6 ± 10.2 | 51.2 ± 4.0 |
| 150 | 83.7 ± 13.5 | 67.2 ± 10.8 | 80.2 ± 15.4 | 63.4 ± 10.8 |
| 175 | 95.7 ± 4.6 | 84.1 ± 11.7 | 94.0 ± 6.3 | 80.6 ± 12.5 |
| 200 | 99.0 ± 1.0 | 89.5 ± 12.5 | 98.0 ± 2.0 | 87.7 ± 11.4 |
| 225 | 98.6 ± 1.6 | 94.2 ± 9.6 | 97.4 ± 3.1 | 91.2 ± 9.9 |
| 250 | 98.5 ± 1.2 | 94.9 ± 6.9 | 96.4 ± 3.2 | 91.0 ± 7.7 |
| 275 | 98.0 ± 1.6 | 96.1 ± 2.8 | 94.8 ± 3.8 | 90.9 ± 5.6 |
| 300 | 97.3 ± 1.8 | 95.1 ± 2.7 | 93.0 ± 4.9 | 88.0 ± 6.0 |
| 325 | 96.7 ± 2.0 | 94.4 ± 2.8 | 91.2 ± 5.1 | 85.3 ± 6.9 |
| 350 | 96.6 ± 1.5 | 94.3 ± 2.6 | 90.0 ± 5.1 | 83.9 ± 6.8 |
| 375 | 96.0 ± 1.7 | 93.5 ± 2.4 | 88.0 ± 5.6 | 81.7 ± 6.9 |
| 400 | 96.4 ± 1.3 | 94.1 ± 2.3 | 88.6 ± 4.5 | 82.0 ± 6.4 |
| 450 | 96.5 ± 1.4 | 94.2 ± 2.0 | 88.6 ± 3.6 | 80.9 ± 5.7 |
| 500 | 97.1 ± 1.0 | 94.7 ± 1.7 | 88.8 ± 3.1 | 80.6 ± 5.6 |
| 550 | 97.1 ± 1.0 | 94.7 ± 1.4 | 89.4 ± 2.5 | 80.6 ± 3.9 |
| 600 | 97.9 ± 0.9 | 95.7 ± 2.0 | 90.4 ± 3.2 | 81.3 ± 6.2 |
| 650 | 98.5 ± 0.8 | 96.0 ± 1.9 | 92.4 ± 2.2 | 82.2 ± 6.7 |
| 700 | 98.8 ± 0.3 | 96.2 ± 1.6 | 93.3 ± 2.3 | 81.3 ± 6.6 |
| 750 | 99.3 ± 0.4 | 96.6 ± 1.7 | 94.3 ± 2.0 | 81.3 ± 7.6 |
| 800 | 99.6 ± 0.2 | 97.2 ± 1.5 | 95.1 ± 3.0 | 82.4 ± 7.8 |
| 850 | 99.7 ± 0.2 | 97.3 ± 2.0 | 96.7 ± 1.7 | 83.0 ± 10.4 |
| 900 | 99.8 ± 0.2 | 97.5 ± 3.4 | 97.2 ± 2.0 | 83.2 ± 11.8 |