Pre-registered training-curve forecast hits: TabPFN covers 14/14 later checkpoints
ZimingLiu11 · x · 2026-10-31
Ziming Liu's team shows their August 28 pre-registered forecast of Paloma macro loss on the Marin Hero Run remained accurate across 14 subsequent checkpoints.
- A pre-trained TabPFN (tabular foundation model) serves as a strong meta-model for scaling-curve prediction with zero extra training, beating NCPL and generalizing to OOD scenarios.
- The TabPFN(L+P) route achieves MAE 0.0216 with a nominal 90% interval covering 14/14 checkpoints, while Marin's official scaling law (MAE 0.0576) offers no uncertainty interval.
Related event: TabPFN's Pre-registered Training Curve Predictions Validated(2 posts)→
More from Models
- Rumor: OpenAI delayed new model over alignment issues, more than any other lab — haider1 · 2026-10-31
- ARC-AGI Kaggle competitions hit major slowdowns as ARC Prize refuses deadline extension — JFPuget · 2026-10-31
- Speech model taxonomy: native S2S and full-duplex are independent dimensions, not exclusive categories — rdesh26 · 2026-10-31
- KAN author argues training-curve prediction fits tabular foundation models like TabPFN, not LLMs — ZimingLiu11 · 2026-10-31
- Code-switched questions hurt LLM accuracy; CANVAS steering fixes internal representations — burkov · 2026-10-31
- OpenAI reportedly to train on users' private sessions with no opt-out — evilsocket · 2026-10-31