#Rrcm

RRCM uses GRPO to learn when to retrieve evidence for LLM recommendation, outperforming fixed-context baselines.