{"id":7429,"date":"2025-09-16T06:19:00","date_gmt":"2025-09-15T21:19:00","guid":{"rendered":"https:\/\/devneko.jp\/wordpress\/?p=7429"},"modified":"2025-09-14T09:21:55","modified_gmt":"2025-09-14T00:21:55","slug":"test-time-scaling-in-reasoning-models-is-not-effective-for-knowledge-intensive-tasks-yet","status":"publish","type":"post","link":"https:\/\/devneko.jp\/wordpress\/?p=7429","title":{"rendered":"Test-Time Scaling in Reasoning Models Is Not Effective for Knowledge-Intensive Tasks Yet"},"content":{"rendered":"\n<ul class=\"wp-block-list\">\n<li><strong>Test-Time Scaling in Reasoning Models Is Not Effective for Knowledge-Intensive Tasks Yet\u00a0<\/strong>[93.0]<br>\u30c6\u30b9\u30c8\u30bf\u30a4\u30e0\u30b9\u30b1\u30fc\u30ea\u30f3\u30b0\u306f\u3001\u30e2\u30c7\u30eb\u304c\u9577\u3044\u63a8\u8ad6\u30c1\u30a7\u30fc\u30f3\u3092\u751f\u6210\u3059\u308b\u3053\u3068\u306b\u3088\u3063\u3066\u3001\u63a8\u8ad6\u6642\u9593\u8a08\u7b97\u3092\u5897\u52a0\u3055\u305b\u308b\u3002 \u672c\u624b\u6cd5\u306f,\u77e5\u8b58\u96c6\u7d04\u578b\u30bf\u30b9\u30af\u306b\u304a\u3044\u3066,\u9ad8\u3044\u4e8b\u5b9f\u7684\u7cbe\u5ea6\u3068\u4f4e\u5e7b\u899a\u7387\u304c\u4e0d\u53ef\u6b20\u3067\u3042\u308b\u5834\u5408\u306b\u304a\u3044\u3066,\u307e\u3060\u6709\u52b9\u3067\u306f\u306a\u3044\u3053\u3068\u3092\u793a\u3059\u3002 \u4ee5\u4e0a\u306e\u7d50\u679c\u304b\u3089,\u30c6\u30b9\u30c8\u6642\u9593\u8a08\u7b97\u306e\u5897\u5927\u306f\u5fc5\u305a\u3057\u3082\u7cbe\u5ea6\u306e\u5411\u4e0a\u306b\u306f\u81f3\u3089\u305a,\u591a\u304f\u306e\u5834\u5408\u306b\u304a\u3044\u3066\u5e7b\u899a\u306e\u5897\u5927\u306b\u3064\u306a\u304c\u308b\u3053\u3068\u304c\u793a\u5506\u3055\u308c\u305f\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2509.06861v1\">\u8ad6\u6587<\/a>\u00a0\u00a0<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2509.06861v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>\u00a0 \u00a0(Mon, 08 Sep 2025 16:28:25 GMT)<\/li>\n\n\n\n<li>\u300cTo summarize, while test-time scaling in reasoning models has led to strong performance in many domains, it is not yet effective for knowledge-intensive tasks. Increasing inference time does not consistently improve factual accuracy, and contrary to expectations, it can even increase hallucinations.\u300d\u3068\u306e\u3053\u3068\u3002LRM\u3092\u4f7f\u3063\u3066\u3044\u3066\u611f\u3058\u3066\u3044\u308b\u3053\u3068\u3068\u6574\u5408\u7684\u3002<\/li>\n\n\n\n<li>\u30ea\u30dd\u30b8\u30c8\u30ea\u306f<a href=\"https:\/\/github.com\/XuZhao0\/tts-knowledge\">GitHub &#8211; XuZhao0\/tts-knowledge: Code and data for &#8220;Test-time scaling in reasoning models is not effective for knowledge-intensive tasks yet&#8221;<\/a><\/li>\n<\/ul>\n","protected":false},"excerpt":{"rendered":"","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[2],"tags":[232,366],"class_list":["post-7429","post","type-post","status-publish","format-standard","hentry","category-arxiv","tag-lrm","tag-slow-thinking"],"_links":{"self":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/7429","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=7429"}],"version-history":[{"count":1,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/7429\/revisions"}],"predecessor-version":[{"id":7430,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/7429\/revisions\/7430"}],"wp:attachment":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=7429"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=7429"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=7429"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}