Safety-critical scenario discovery on UTM Out-of-distribution Region TR2
100APO (%)Smoke Test
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Smoke TestMethod Type=Basic functionality testing2026.06 | 100 | — | — | — | — | 1 | |
| Expert-Guided ExploitationMethod Type=Expert testing2026.06 | 83.3 | 25 | 1.7 | 3.9 | 5.8 | 1 | |
| PM-1.2BModel Scale=1.2B, Method Type=Policy Model2026.06 | 38.3 | 30 | 4.2 | 4.5 | 17.6 | 2.2 | |
| PM-2BModel Scale=2B, Method Type=Policy Model2026.06 | 31.5 | 46 | 4.9 | 64.1 | 50.5 | 7.6 |