Legal Question Answering on LEXam Public Law (n=12)
83.3AccuracyZS (no traces/HUBO)
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| ZS (no traces/HUBO)Adjudicator=Opus 4.62026.06 | 83.3 | — | |
| EP-HUBO (Qwen-35B)Adjudicator=Opus 4.6, Trace Generator=Qwen3.5-35B2026.06 | 66.7 | -16.7 | |
| EP-HUBO (OSS-20B)Adjudicator=Opus 4.6, Trace Generator=OSS-20B2026.06 | 66.7 | -16.7 | |
| MV (Qwen-35B)Trace Generator=Qwen3.5-35B2026.06 | 58.3 | — | |
| ZS (no traces/HUBO)Adjudicator=Sonnet 4.62026.06 | 50 | — | |
| EP-HUBO (Qwen-35B)Adjudicator=Sonnet 4.6, Trace Generator=Qwen3.5-35B2026.06 | 50 | 0 | |
| EP-HUBO (OSS-20B)Adjudicator=Sonnet 4.6, Trace Generator=OSS-20B2026.06 | 33.3 | -16.7 | |
| MV (OSS-20B)Trace Generator=OSS-20B2026.06 | 25 | — |