SPoT generalizes out-of-distribution: models learn to flag unanswerable questions
polkirichenko · x · 2026-11-12
Part 5 of the SPoT thread shows a striking generalization result: despite training only on answerable math problems, SPoT improves models' ability to abstain on unanswerable questions, generalizing outside the math domain (clear AUROC gains on out-of-domain sets like BBQ, MoralChoice, GPQA-D, MediQ).
In the example, the original reasoning chain forces an answer of 9, while post-thinking catches the incorrect assumption (the problem never states the total number of trees), identifies the question as unanswerable, and assigns low confidence (0.1) to the original answer.
Related event: SPoT calibration boosts answerability detection(2 posts)→
More from Research
- As models get stronger, the next trainable tasks look increasingly like alignment itself — herbiebradley · 2026-11-12
- RoboCap-10k: 10,000 hours of open egocentric robot data across 21 industries released — micoolcho · 2026-11-12
- Talent vs. luck model reproduced: equal funding is ~9x more efficient than funding winners — ChenhaoTan · 2026-11-12
- Stanford HAI dashboards show employment in AI-exposed jobs down over 4% this year — StanfordHAI · 2026-11-12
- Shun-ichi Amari on how 'do as you like' freedom birthed the math behind AI's 60 years — FrnkNlsn · 2026-11-12
- Formal verification proves a 3,000-step Sokoban optimum, with a bounty on beating it — banteg · 2026-11-12