{"type":"link","version":"1.0","title":"Agentic RL training recovers the citation precision that SFT sacrifices: O-Researcher improves Citation Accuracy 29.13%→31.99% via RL after SFT while outperforming commercial systems including O3-Deep-Research on retrieval faithfulness","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/agentic-rl-recovers-sft-precision-loss","thumbnail_url":"https://ai-arch.pages.dev/android-chrome-512x512.png","thumbnail_width":512,"thumbnail_height":512}