{"id":3488,"date":"2023-06-27T06:24:00","date_gmt":"2023-06-26T21:24:00","guid":{"rendered":"https:\/\/devneko.jp\/wordpress\/?p=3488"},"modified":"2023-06-27T06:24:00","modified_gmt":"2023-06-26T21:24:00","slug":"gpt-4%e3%81%a8gpt-3-5%e3%81%ae%e4%bf%a1%e9%a0%bc%e6%80%a7","status":"publish","type":"post","link":"https:\/\/devneko.jp\/wordpress\/?p=3488","title":{"rendered":"GPT-4\u3068GPT-3.5\u306e\u4fe1\u983c\u6027"},"content":{"rendered":"\n<ul class=\"wp-block-list\">\n<li><strong>DecodingTrust: A Comprehensive Assessment of Trustworthiness in GPT Models\u00a0<\/strong>[76.8]<br>\u672c\u7a3f\u3067\u306f,GPT-4\u3068GPT-3.5\u306b\u7740\u76ee\u3057\u305f\u5927\u898f\u6a21\u8a00\u8a9e\u30e2\u30c7\u30eb\u306e\u7dcf\u5408\u7684\u4fe1\u983c\u6027\u8a55\u4fa1\u3092\u63d0\u6848\u3059\u308b\u3002 \u8a55\u4fa1\u306e\u7d50\u679c,\u4fe1\u983c\u611f\u306e\u8105\u5a01\u306b\u5bfe\u3059\u308b\u672a\u516c\u8868\u306e\u8106\u5f31\u6027\u304c\u5224\u660e\u3057\u305f\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2306.11698v1\">\u8ad6\u6587<\/a>\u00a0\u00a0<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2306.11698v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>\u00a0 \u00a0(Tue, 20 Jun 2023 17:24:23 GMT)<\/li>\n\n\n\n<li>GPT-4\u3068GPT-3.5\u306e\u4fe1\u983c\u6027\u3092\u691c\u8a3c\u3057\u305f\u8ad6\u6587\u3002\u901a\u5e38\u306fGPT-4\u306e\u65b9\u304c\u4fe1\u983c\u6027\u304c\u9ad8\u3044\u304c\u300cGPT-4 is more vulnerable given jailbreaking system or user prompts\u300d\u3068\u306e\u3053\u3068\u3002GPT-4\u306f\uff08jailbreaking\u3055\u308c\u305f\u5834\u5408\u3082\uff09\u3088\u308a\u5fe0\u5b9f\u306b\u547d\u4ee4\u306b\u5f93\u304a\u3046\u3068\u3059\u308b\u305f\u3081\u304b\u3082\u3057\u308c\u306a\u3044\u3068\u3057\u3066\u3044\u308b\u300290\u30da\u30fc\u30b8\u3068\u9577\u3044\u304c\u3001\u975e\u5e38\u306b\u8a73\u7d30\u306a\u691c\u8a3c\u304c\u306a\u3055\u308c\u3066\u3044\u3066\u3068\u3066\u3082\u52c9\u5f37\u306b\u306a\u308b\u3002<\/li>\n\n\n\n<li>\u30d7\u30ed\u30b8\u30a7\u30af\u30c8\u30b5\u30a4\u30c8\u306f<a href=\"https:\/\/decodingtrust.github.io\/\">DecodingTrust Benchmark<\/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":[173,175,387,622],"class_list":["post-3488","post","type-post","status-publish","format-standard","hentry","category-arxiv","tag-gpt","tag-gpt-4","tag-survey","tag-622"],"_links":{"self":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/3488","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=3488"}],"version-history":[{"count":0,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/3488\/revisions"}],"wp:attachment":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=3488"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=3488"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=3488"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}