[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"topic-en-pagedweight-trims-moe-memory-without-tanking-quality":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},"0047cc2e-a579-45b8-a937-5764d6665393","pagedweight-trims-moe-memory-without-tanking-quality","PagedWeight trims MoE memory without tanking quality",false,"active","PagedWeight dynamically quantizes MoE weights at runtime to trade GPU memory for KV cache headroom.","research","en","2026-07-20T06:02:28.360603+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},"89e53d67-0144-4862-ad09-71f834b878f4","pagedweight-moe-serving-dynamic-quantization-en","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1784527372051-foet.png","2026-07-20T06:02:27.566+00:00",true]