{"id":3992,"date":"2023-10-30T05:23:00","date_gmt":"2023-10-29T20:23:00","guid":{"rendered":"https:\/\/devneko.jp\/wordpress\/?p=3992"},"modified":"2023-10-30T05:23:00","modified_gmt":"2023-10-29T20:23:00","slug":"language-models-hallucinate-but-may-excel-at-fact-verification","status":"publish","type":"post","link":"https:\/\/devneko.jp\/wordpress\/?p=3992","title":{"rendered":"Language Models Hallucinate, but May Excel at Fact Verification"},"content":{"rendered":"\n<ul class=\"wp-block-list\">\n<li><strong>Language Models Hallucinate, but May Excel at Fact Verification\u00a0<\/strong>[95.6]<br>\u5927\u898f\u6a21\u8a00\u8a9e\u30e2\u30c7\u30eb(LLM)\u306f\u3057\u3070\u3057\u3070\u300c\u5e7b\u60d1(hallucinate)\u300d\u3057\u3001\u7d50\u679c\u3068\u3057\u3066\u975e\u5b9f\u52b9\u51fa\u529b\u3068\u306a\u308b\u3002 GPT-3.5\u3067\u3055\u3048\u3001\u4e8b\u5b9f\u306e\u51fa\u529b\u306f25%\u4ee5\u4e0b\u3067\u3042\u308b\u3002 \u3053\u308c\u306f\u3001\u9032\u6357\u3092\u8a08\u6e2c\u3057\u3001\u30a4\u30f3\u30bb\u30f3\u30c6\u30a3\u30d6\u3092\u4e0e\u3048\u308b\u305f\u3081\u306b\u3001\u4e8b\u5b9f\u691c\u8a3c\u306e\u91cd\u8981\u6027\u3092\u6d6e\u304d\u5f6b\u308a\u306b\u3059\u308b\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2310.14564v1\">\u8ad6\u6587<\/a>\u00a0\u00a0<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2310.14564v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>\u00a0 \u00a0(Mon, 23 Oct 2023 04:39:01 GMT)<\/li>\n\n\n\n<li>\u69d8\u3005\u306a\u30e2\u30c7\u30eb\u3084\u6761\u4ef6\u3067\u306e\u30cf\u30eb\u30b7\u30cd\u30fc\u30b7\u30e7\u30f3\u306b\u3064\u3044\u3066\u691c\u8a3c\u3057\u305f\u8ad6\u6587\u3002\u300cSurprisingly, FLAN-T511B, the least factual generator in our study, performs the best as a fact verifier, even outperforming more capable LLMs like GPT3.5 and ChatGPT.\u300d\u3068\u3044\u3046\u7d50\u679c\u3002<\/li>\n\n\n\n<li>\u300cThe overall inferior performance of not using evidence reveals the importance of retrieval.\u300d\u306f\u500b\u4eba\u7684\u306a\u611f\u899a\u306b\u3082\u3042\u3046\u3002<\/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":[184,223],"class_list":["post-3992","post","type-post","status-publish","format-standard","hentry","category-arxiv","tag-hallucination","tag-llm"],"_links":{"self":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/3992","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=3992"}],"version-history":[{"count":0,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/3992\/revisions"}],"wp:attachment":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=3992"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=3992"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=3992"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}