COVA-SFT dataset trains models to solve text problems with visual chain-of-thought
yoavartzi · x · 2026-10-16
- Researchers released COVA-SFT, a large-scale public dataset for teaching models to solve text-based problems via visual chain-of-thought, based on the idea that humans think visually and abstractly even for text-only problems.
- The work will be presented at BayLearn 2026 and accepted to EMNLP 2026 Findings.
- The team notes multiple groups are now exploring this direction; the paper and dataset are publicly available.
More from Research
- Free Textbook: Introduction to Machine Learning Covers the Math from Optimization to Generative Models — TheTuringPost · 2026-10-16
- Free ML math textbook: Laurent Younes' Introduction to Machine Learning covers Adam to generative models — TheTuringPost · 2026-10-16
- RFdiffusion3 published in Science, alongside Chaperone agent harness for bio-molecular design — anshulkundaje · 2026-10-16
- Microsoft paper: history-aware model routing lifts multi-turn agent task pass rate by 3.3 points — dair_ai · 2026-10-16
- Researcher: everyone should build a good eval at least once — bad evals are everywhere — gabriberton · 2026-10-16
- The square lattice connective constant remains unsolved — and might be ripe for AI — tak3sh8 · 2026-10-16