phi3star
done · peak unseen_mismatch 93.1
Referenced by
α schedule
α 3.1993 ×25 α 3.155 ×25 α 3.0753 ×25 α 2.963 ×25 α 2.8198 ×25 α 2.6469 ×25 α 2.4453 ×25 α 2.2158 ×25 α 1.9588 ×25 α 1.675 ×25 α 1.3649 ×25 α 1.0288 ×25 α 0.6671 ×25 α 0.2802 ×25 α 0.2802 ×50
Figures
Results
accuracy by training iteration, mean ± sd over seeds · schedule ends at α = 0.2802
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 |
|---|---|---|---|---|
| 100 | 56.4 ± 6.6 | 52.8 ± 4.1 | 52.8 ± 5.8 | 50.5 ± 3.1 |
| 110 | 59.8 ± 9.1 | 55.5 ± 6.5 | 55.1 ± 8.3 | 51.9 ± 4.8 |
| 120 | 64.3 ± 11.2 | 59.7 ± 8.8 | 58.2 ± 10.5 | 54.3 ± 6.8 |
| 130 | 70.0 ± 12.4 | 64.8 ± 10.7 | 62.6 ± 12.4 | 57.4 ± 8.8 |
| 140 | 75.0 ± 12.8 | 69.8 ± 11.5 | 66.5 ± 13.7 | 61.1 ± 10.8 |
| 150 | 79.6 ± 11.9 | 74.8 ± 10.7 | 70.4 ± 13.7 | 65.0 ± 11.1 |
| 160 | 83.5 ± 11.1 | 79.3 ± 10.5 | 73.9 ± 13.4 | 69.0 ± 11.7 |
| 170 | 86.6 ± 9.3 | 82.7 ± 9.2 | 76.7 ± 12.6 | 71.9 ± 11.3 |
| 180 | 88.9 ± 7.5 | 85.5 ± 7.7 | 79.0 ± 11.2 | 74.3 ± 10.0 |
| 190 | 90.9 ± 6.3 | 87.9 ± 6.4 | 81.2 ± 10.0 | 76.6 ± 9.1 |
| 200 | 92.2 ± 5.3 | 89.6 ± 5.3 | 82.5 ± 8.8 | 78.2 ± 8.4 |
| 210 | 93.4 ± 4.7 | 90.9 ± 4.6 | 83.7 ± 8.2 | 79.6 ± 7.8 |
| 220 | 94.3 ± 4.0 | 92.2 ± 4.2 | 84.7 ± 7.3 | 80.8 ± 7.1 |
| 230 | 95.2 ± 3.5 | 93.3 ± 3.6 | 86.4 ± 6.7 | 82.6 ± 6.7 |
| 240 | 96.4 ± 2.5 | 94.7 ± 2.6 | 88.7 ± 5.9 | 85.2 ± 5.8 |
| 250 | 96.8 ± 2.1 | 95.2 ± 2.2 | 89.1 ± 4.6 | 85.5 ± 4.6 |
| 260 | 97.8 ± 1.7 | 96.6 ± 1.9 | 91.8 ± 4.1 | 88.3 ± 4.6 |
| 270 | 98.4 ± 1.5 | 97.2 ± 1.7 | 93.7 ± 3.5 | 89.6 ± 4.4 |
| 280 | 98.8 ± 1.2 | 97.5 ± 2.1 | 94.6 ± 2.9 | 89.8 ± 4.7 |
| 290 | 99.5 ± 0.5 | 98.4 ± 1.0 | 97.0 ± 1.7 | 92.1 ± 3.0 |
| 300 | 99.9 ± 0.3 | 99.0 ± 0.4 | 98.1 ± 1.6 | 93.1 ± 2.8 |
| 310 | 100.0 ± 0.0 | 98.7 ± 1.0 | 99.0 ± 0.2 | 88.9 ± 7.1 |
| 320 | 100.0 ± 0.0 | 98.7 ± 0.3 | 99.7 ± 0.1 | 85.5 ± 3.8 |
| 330 | 100.0 ± 0.0 | 98.9 ± 0.3 | 99.9 ± 0.1 | 89.2 ± 0.1 |
| 340 | 100.0 ± 0.1 | 99.0 ± 0.3 | 100.0 ± 0.1 | 91.9 ± 0.7 |