[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"tag-semantic-evaluation":3},{"tag":4,"articles":10},{"id":5,"name":6,"slug":7,"article_count":8,"description_zh":9,"description_en":9},"90acbe8f-dac5-4483-8070-1b06cc76e25a","semantic evaluation","semantic-evaluation",1,null,[11],{"id":12,"slug":13,"title":14,"summary":15,"category":16,"image_url":17,"cover_image":17,"language":18,"created_at":19},"b41b3999-fa8c-4e87-8914-4ed027fe8bfe","llms-for-asr-evaluation-beyond-wer-zh","LLM 評測 ASR 不只看 WER","這篇論文把 decoder-based LLM 拿來當 ASR 評測器，結果在人工一致性上明顯贏過 WER；在 HATS 的二選一任務，最佳模型達 92–94%。","research","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1777010993292-cy1y.png","zh","2026-04-24T06:09:37.70822+00:00"]