Relevance Correlation Evaluation on HJM (Equal)
0.525Spearman rhoPlain LLM judge
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Plain LLM judgeJudge Type=Plain2026.07 | 0.525 | 0.486 | |
| Behavior-Grounded LLM judgeJudge Type=Behavior-Grounded2026.07 | 0.525 | 0.486 |
Loading the SOTA2 catalog…