[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"tag-prerl":3},{"tag":4,"articles":10},{"id":5,"name":6,"slug":7,"article_count":8,"description_zh":9,"description_en":9},"437745a9-1a74-4949-b279-5510d5c1d885","PreRL","prerl",1,null,[11],{"id":12,"slug":13,"title":14,"summary":15,"category":16,"image_url":17,"cover_image":17,"language":18,"created_at":19},"2428c4f3-8cbf-43dc-afe8-dad89550740f","prerl-training-llms-in-pre-train-space-zh","PreRL：把強化學習搬進預訓練空間","PreRL 把 RL 從 P(y|x) 轉向 P(y)，直接在預訓練空間做獎勵更新，主打增強推理與探索。摘要也提到 NSR 與 DSRL 兩種設計。","research","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1776319619099-op5n.png","zh","2026-04-16T06:06:37.875971+00:00"]