Answer Verification on OlympiadBench terminal answers
0.9546AUROCMSV 64
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| MSV 64N sequences=64, Weighted Voting=False2026.03 | 0.9546 | 0.0835 | |
| MSV 16N sequences=16, Weighted Voting=False2026.03 | 0.9469 | 0.1005 | |
| MSV 4N sequences=4, Weighted Voting=False2026.03 | 0.9298 | 0.1108 | |
| Probe + WV 64N sequences=64, Weighted Voting=True2026.03 | 0.9173 | 0.1466 | |
| MSV 1 + WV 64N sequences=64, Weighted Voting=True2026.03 | 0.913 | 0.1499 | |
| Probe + WV 16N sequences=16, Weighted Voting=True2026.03 | 0.9028 | 0.1588 | |
| ProbeN sequences=1, Weighted Voting=False2026.03 | 0.9022 | 0.1271 | |
| MSV 1 + WV 16N sequences=16, Weighted Voting=True2026.03 | 0.9008 | 0.1612 | |
| MSV 1N sequences=1, Weighted Voting=False2026.03 | 0.8926 | 0.131 | |
| Probe + WV 4N sequences=4, Weighted Voting=True2026.03 | 0.8494 | 0.2164 | |
| MSV 1 + WV 4N sequences=4, Weighted Voting=True2026.03 | 0.8479 | 0.2168 |