[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"tag-leaderboards":3},{"tag":4,"articles":9},{"id":5,"name":6,"slug":6,"article_count":7,"description_zh":8,"description_en":8},"e87550bb-6fdd-4f07-b2a3-6e37e4988e49","leaderboards",0,null,[10],{"id":11,"slug":12,"title":13,"summary":14,"category":15,"image_url":16,"cover_image":16,"language":17,"created_at":18},"1848b0d4-2c8a-4c24-928b-46f0ddb4dbb2","why-benchmark-leaderboards-are-wrong-about-model-logic-en","Why benchmark leaderboards are wrong about model logic","Leaderboard churn overstates progress and hides how weak model logic still is.","research","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1780673573292-rj31.png","en","2026-06-05T15:32:23.511842+00:00"]