AI RESEARCH
x1: Learning to Think Adaptively Across Languages and Cultures
arXiv CS.CL
•
ArXi:2604.16917v1 Announce Type: new Languages encode distinct abstractions and inductive priors, yet most large language models (LLMs) overlook this diversity by reasoning in a single dominant language. In this work, we