| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| FACTOR | HICD | Factual Consistency (Wiki)64.43 | 14 | 3mo ago | |
| RecRM-Bench | Ours | Accuracy (%)70.71 | 8 | 21d ago | |
| Noisy-Concept-Captions Denoised Latent, iterations 21-30 | NoisyCLIP | R@150.6 | 4 | 3mo ago | |
| POPE | MAGIC | POPE Score84.8 | 3 | 8d ago |