{"id":7969,"date":"2025-12-24T17:43:00","date_gmt":"2025-12-24T08:43:00","guid":{"rendered":"https:\/\/devneko.jp\/wordpress\/?p=7969"},"modified":"2025-12-21T21:47:18","modified_gmt":"2025-12-21T12:47:18","slug":"are-we-on-the-right-way-to-assessing-llm-as-a-judge","status":"publish","type":"post","link":"https:\/\/devneko.jp\/wordpress\/?p=7969","title":{"rendered":"Are We on the Right Way to Assessing LLM-as-a-Judge?\u00a0"},"content":{"rendered":"\n<ul class=\"wp-block-list\">\n<li><strong>Are We on the Right Way to Assessing LLM-as-a-Judge?\u00a0<\/strong>[16.3]<br>\u4eba\u9593\u306e\u30a2\u30ce\u30c6\u30fc\u30b7\u30e7\u30f3\u3092\u5fc5\u8981\u3068\u305b\u305a\u306bLCM\u5be9\u67fb\u54e1\u306e\u8cea\u3092\u8a55\u4fa1\u3059\u308b\u65b0\u3057\u3044\u8a55\u4fa1\u30b9\u30a4\u30fc\u30c8\u3067\u3042\u308bSage\u3092\u7d39\u4ecb\u3059\u308b\u3002 \u5408\u7406\u7684\u9078\u629e\u7406\u8ad6\u306e\u516c\u7406\u306b\u89e6\u767a\u3055\u308c\u305f\u30bb\u30fc\u30b8\u306f\u3001LLM-as-a-Judge\u3092\u6e2c\u5b9a\u3059\u308b\u305f\u3081\u306e2\u3064\u306e\u65b0\u3057\u3044\u30ec\u30f3\u30ba(\u5c40\u6240\u7684\u306a\u81ea\u5df1\u6574\u5408\u6027\u3068\u5927\u57df\u7684\u306a\u8ad6\u7406\u7684\u6574\u5408\u6027)\u3092\u5c0e\u5165\u3057\u305f\u3002 Sage \u306b\u57fa\u3065\u3044\u3066,\u73fe\u5728\u6700\u5148\u7aef\u306e LLM \u306f,\u30b9\u30b3\u30a2\u30ea\u30f3\u30b0\u3068\u30da\u30a2\u30ef\u30a4\u30ba\u306e\u4e21\u65b9\u306b\u304a\u3044\u3066,\u5be9\u67fb\u54e1\u3068\u3057\u3066\u6a5f\u80fd\u3059\u308b\u5834\u5408\u306b,\u91cd\u5927\u306a\u4fe1\u983c\u6027\u4e0a\u306e\u554f\u984c\u304c\u3042\u308b\u3053\u3068\u3092\u660e\u3089\u304b\u306b\u3057\u305f\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2512.16041v1\">\u8ad6\u6587<\/a>\u00a0\u00a0<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2512.16041v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>\u00a0 \u00a0(Wed, 17 Dec 2025 23:49:55 GMT)<\/li>\n\n\n\n<li>LLM-as-a-Judge\u3092\u8a55\u4fa1\u3059\u308b\u30d9\u30f3\u30c1\u30de\u30fc\u30af\u306e\u63d0\u6848\u3001\u304a\u3088\u3073\u300cOur experiments reveal significant robustness deficiencies in current state-of-the-art models. We attribute these inconsistent judgments to a newly identified phenomenon called situational preference where models fail to maintain a stable internal gauging principle across different contexts.  To address this, we demonstrate that implementing self-generated rubrics effectively mitigates situational preference and boosts judgment consistency. We also investigate the impact of fine-tuning and explanatory reasoning on evaluation performance.\u300d\u3068\u306e\u6307\u6458\u3002<\/li>\n\n\n\n<li>\u30ea\u30dd\u30b8\u30c8\u30ea\u306f<a href=\"https:\/\/entroplay.ai\/research\/rethinking-llm-as-a-judge\">Entroplay.ai<\/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":[224],"class_list":["post-7969","post","type-post","status-publish","format-standard","hentry","category-arxiv","tag-llm-as-a-judge"],"_links":{"self":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/7969","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=7969"}],"version-history":[{"count":1,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/7969\/revisions"}],"predecessor-version":[{"id":7970,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/7969\/revisions\/7970"}],"wp:attachment":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=7969"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=7969"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=7969"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}