r8_c0
done · peak unseen_mismatch 96.8
α schedule
α 3.7185 ×25 α 3.7149 ×25 α 3.7007 ×25 α 1.4718 ×25 α 1.4482 ×25 α 3.5157 ×25 α 3.3796 ×25 α 3.1924 ×25 α 1.182 ×25 α 1.0556 ×25 α 0.8985 ×25 α 1.7624 ×25 α 1.1913 ×25 α 0.2076 ×25
Figures
Results
accuracy by training iteration, mean ± sd over seeds · schedule ends at α = 0.2076
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 | 50.7 ± 1.3 | 50.3 ± 1.7 | 49.9 ± 1.7 | 49.5 ± 1.9 |
| 50 | 51.4 ± 1.6 | 50.3 ± 1.6 | 50.0 ± 1.2 | 49.8 ± 1.8 |
| 75 | 52.2 ± 2.2 | 50.0 ± 1.7 | 50.5 ± 1.3 | 49.4 ± 1.5 |
| 100 | 54.6 ± 5.5 | 51.9 ± 2.3 | 51.8 ± 3.1 | 49.8 ± 1.5 |
| 125 | 64.6 ± 14.7 | 55.4 ± 6.7 | 60.5 ± 14.5 | 52.3 ± 6.1 |
| 150 | 77.6 ± 16.5 | 71.0 ± 13.6 | 73.6 ± 17.7 | 65.9 ± 13.5 |
| 175 | 87.3 ± 10.8 | 82.5 ± 10.7 | 80.5 ± 13.7 | 75.2 ± 12.2 |
| 200 | 90.7 ± 6.3 | 88.0 ± 6.3 | 82.0 ± 10.1 | 78.7 ± 9.9 |
| 225 | 98.7 ± 2.4 | 96.6 ± 3.5 | 97.3 ± 3.2 | 93.8 ± 5.6 |
| 250 | 99.8 ± 0.5 | 97.1 ± 5.9 | 99.0 ± 1.1 | 94.8 ± 6.4 |
| 275 | 100.0 ± 0.1 | 98.1 ± 3.3 | 99.9 ± 0.3 | 96.6 ± 4.7 |
| 300 | 99.0 ± 0.9 | 97.5 ± 1.4 | 95.8 ± 4.0 | 91.9 ± 5.6 |
| 325 | 99.8 ± 0.2 | 99.1 ± 0.9 | 99.2 ± 1.1 | 96.8 ± 3.3 |
| 350 | 100.0 ± 0.0 | 97.3 ± 7.9 | 100.0 ± 0.0 | 96.7 ± 9.0 |