[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"topic-zh-topic-c115d203":3},{"cluster":4,"timeline":16},{"id":5,"slug":6,"title":7,"pinned":8,"status":9,"summary":10,"category":11,"language":12,"created_at":13,"merged_into":14,"article_count":15,"first_seen_at":13,"last_updated_at":13},"f9604e6c-c45f-40be-99d2-0a53c5d7d02c","topic-c115d203","安全持續學習還沒解題",false,"active","這篇 arXiv 研究把安全 RL 和持續 RL 放在一起看，指出環境一變，現有方法常常顧不了安全，也守不住舊行為。","research","zh","2026-07-05T09:04:36.112684+00:00",null,1,[17],{"id":18,"slug":19,"title":7,"summary":10,"category":11,"image_url":20,"cover_image":20,"published_at":21,"is_canonical_seed":22},"947e3be0-2b4b-4719-90d1-ddd1ac80f18a","safe-continual-rl-changing-environments-zh","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1776838196623-anqk.png","2026-04-22T06:09:32.378+00:00",true]