Cohere Labs: data mixing enables in-language reasoning without multilingual supervision
CohereLabs · hf · 2026-09-12
Cohere Labs released a study showing optimized SFT data composition lets reasoning models consistently process and respond in diverse non-English languages without requiring reasoning supervision in each target language. Data mixing acts as the pillar of cross-lingual generalization, cutting multilingual training data costs.
Related event: Cohere Open-Sources Tiny Aya L2-Thinker for Native Multilingual Reasoning(4 posts)→
More from Research
- Qwen3.8-Flash-Next's open 180B weights hailed as a model organism for backdoor research — jd_pressman · 2026-10-27
- Latent-Foresight: world models that learn to predict their own future — abursuc · 2026-10-27
- UK AISI papers: no good safeguards yet against misuse of biological AI tools — davidmanheim · 2026-10-27
- CMU's XYBench: LLMs answer queries literally and reinforce user misconceptions — valentina__py · 2026-10-27
- arXiv advice from T. Dietterich: wrong primary category will almost surely get your paper on hold — tdietterich · 2026-10-27
- arXiv has 4,000+ papers on hold amid submission surge; moderator shares how to avoid it — tdietterich · 2026-10-27