{"type":"link","version":"1.0","title":"In retrieval-in-the-loop RL, exclude the retrieved passages from the policy-gradient loss","author_name":"AI Archs","author_url":"https://ai-arch.pages.dev","provider_name":"AI Archs","provider_url":"https://ai-arch.pages.dev","url":"https://ai-arch.pages.dev/n/mask-externally-retrieved-tokens-from-the-rl-policy-loss","thumbnail_url":"https://ai-arch.pages.dev/og/mask-externally-retrieved-tokens-from-the-rl-policy-loss.png","thumbnail_width":1200,"thumbnail_height":630}