Legal Question Answering on LEXam Private Law (n=63)
87.3AccuracyMV (Qwen-35B)
Evaluation results
| Method | Links | ||
|---|---|---|---|
| MV (Qwen-35B)Trace Generator=Qwen3.5-35B2026.06 | 87.3 | — | |
| EP-HUBO (Qwen-35B)Adjudicator=Opus 4.6, Trace Generator=Qwen3.5-35B2026.06 | 87.3 | 1.6 | |
| EP-HUBO (OSS-20B)Adjudicator=Sonnet 4.6, Trace Generator=OSS-20B2026.06 | 87.3 | 11.1 | |
| ZS (no traces/HUBO)Adjudicator=Opus 4.62026.06 | 85.7 | — | |
| EP-HUBO (OSS-20B)Adjudicator=Opus 4.6, Trace Generator=OSS-20B2026.06 | 85.7 | 0 | |
| EP-HUBO (Qwen-35B)Adjudicator=Sonnet 4.6, Trace Generator=Qwen3.5-35B2026.06 | 84.1 | 7.9 | |
| ZS (no traces/HUBO)Adjudicator=Sonnet 4.62026.06 | 76.2 | — | |
| MV (OSS-20B)Trace Generator=OSS-20B2026.06 | 76.2 | — |