| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| Gumbel + Gaussian (alpha=2.0) | Log-FM | Catastrophic Failure Rate0 | 20 | 14d ago | |
| Gumbel + Gaussian alpha=1.5 | Log-FM | Catastrophic Failure Fraction (WP1 > 1)2 | 20 | 14d ago | |
| Spot the Cow | RFM | NLL1.03 | 9 | 2mo ago | |
| Stanford Bunny | RFM | NLL1.22 | 9 | 2mo ago | |
| GridWorld (test) | VRAdam | Flow Matching Loss1.33 | 5 | 21d ago | |
| GridWorld (val) | VRAdam | Flow Matching Loss1.25 | 5 | 21d ago |