[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"tag-recursive-computation":3},{"tag":4,"articles":11},{"id":5,"name":6,"slug":7,"article_count":8,"description_zh":9,"description_en":10},"a3712746-109d-4518-b7b4-c1b67dbad5c7","recursive computation","recursive-computation",3,"遞迴計算指的是把同一組模型或模組反覆套用在內部表示上，以較小的參數與記憶體成本換取更深的推理路徑。這類方法常見於多代理協作、語言模型加深與注意力效率優化，重點在於控制 token、cache 與延遲。","Recursive computation reuses the same model or module across repeated passes over internal representations, trading extra depth for lower parameter, token, and memory costs. It shows up in multi-agent setups, deeper LLM inference, and attention designs that try to limit KV cache growth.",[12],{"id":13,"slug":14,"title":15,"summary":16,"category":17,"image_url":18,"cover_image":18,"language":19,"created_at":20},"27f0d044-b9f9-4a58-99e8-1a181ea32f19","universal-yoco-efficient-depth-scaling-en","Universal YOCO aims to scale depth without cache bloat","YOCO-U mixes recursive computation with efficient attention to scale LLM depth while keeping inference overhead and KV cache growth in check.","research","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1775115621645-wqql.png","en","2026-04-02T06:06:26.960639+00:00"]