Coverage-based Alignment on ICLR 50 submissions 2026
88.6Str-CovPaperAudit
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| PaperAuditreviewer=GPT-5, judge=Gemini-2.5-Pro2026.01 | 88.6 | 59.1 | 46.8 | 44.4 | |
| Baselinereviewer=GPT-5, judge=Gemini-2.5-Pro2026.01 | 87.6 | 56.8 | 49.5 | 43.3 | |
| DeepReviewreviewer=GPT-5, judge=Gemini-2.5-Pro2026.01 | 84.4 | 48.2 | 48.8 | 41.2 |