AI RESEARCH

Reasoning Cache: Continual Improvement Over Long Horizons via Short-Horizon RL

arXiv CS.LG

ArXi:2602.03773v2 Announce Type: replace Large Language Models (LLMs) that can continually improve beyond their