[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"tag-zero-shot-generalization":3},{"tag":4,"articles":10,"peer_article_count":8},{"id":5,"name":6,"slug":7,"article_count":8,"description_zh":9,"description_en":9},"28646460-cb02-4b80-863b-25505ee186f9","zero-shot generalization","zero-shot-generalization",1,null,[11,20],{"id":12,"slug":13,"title":14,"summary":15,"category":16,"image_url":17,"cover_image":17,"language":18,"created_at":19},"5c3cb90f-7efd-426f-8c09-32a303f82be9","humanoid-gpt-zero-shot-motion-tracking-zh","Humanoid-GPT：用 GPT 擴大動作追蹤","Humanoid-GPT 用 GPT 風格 Transformer 與 20 億幀動作資料，主打零樣本全身動作追蹤與更強泛化。","research","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1780469319284-znpc.png","zh","2026-06-03T06:47:34.463464+00:00",{"id":21,"slug":22,"title":23,"summary":24,"category":16,"image_url":25,"cover_image":25,"language":18,"created_at":26},"99ae4e8c-3fb2-4d46-8965-458bd16372c8","steerable-vi-vision-representations-zh","可被文字操控的 ViT 特徵","這篇論文把文字提示塞進 ViT 編碼過程中，讓影像特徵能朝指定物件偏移，同時盡量保留通用視覺用途。","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1775196412799-wufk.png","2026-04-03T06:06:34.443487+00:00"]