AI RESEARCH

From Meta-Thought to Execution: Cognitively Aligned Post-Training for Generalizable and Reliable LLM Reasoning

arXiv CS.AI

ArXi:2601.21909v2 Announce Type: replace Current LLM post-