agentsfactneutralOne-step and two-step methods for policy learning with RAG can be formulated under the potential outcome framework with bounded regretMachine Learning (Statistics)28 Jul 2026http://arxiv.org/abs/2607.18225v1