{"id":6132,"date":"2025-01-28T02:12:00","date_gmt":"2025-01-27T17:12:00","guid":{"rendered":"https:\/\/devneko.jp\/wordpress\/?p=6132"},"modified":"2025-01-28T02:12:00","modified_gmt":"2025-01-27T17:12:00","slug":"debate-helps-weak-to-strong-generalization","status":"publish","type":"post","link":"https:\/\/devneko.jp\/wordpress\/?p=6132","title":{"rendered":"Debate Helps Weak-to-Strong Generalization"},"content":{"rendered":"\n<ul class=\"wp-block-list\">\n<li><strong>Debate Helps Weak-to-Strong Generalization\u00a0<\/strong>[68.7]<br>\u6211\u3005\u306f,\u5f37\u3044\u4e8b\u524d\u8a13\u7df4\u30e2\u30c7\u30eb\u3092\u7528\u3044\u3066\u4eba\u9593\u306e\u76e3\u7763\u3092\u6539\u5584\u3059\u308b\u65b9\u6cd5\u306b\u3064\u3044\u3066\u691c\u8a0e\u3057,\u5f31\u3044\u4eba\u9593\u306e\u76e3\u7763\u3092\u5897\u5f37\u3057\u305f\u5f37\u3044\u30e2\u30c7\u30eb\u3092\u76e3\u7763\u3059\u308b\u3002 \u8b70\u8ad6\u306f\u3001\u4fe1\u983c\u3067\u304d\u306a\u3044\u5f37\u529b\u306a\u30e2\u30c7\u30eb\u304b\u3089\u4fe1\u983c\u3067\u304d\u308b\u60c5\u5831\u3092\u62bd\u51fa\u3059\u308b\u5f31\u3044\u30e2\u30c7\u30eb\u3092\u652f\u63f4\u3059\u308b\u3053\u3068\u304c\u3067\u304d\u308b\u3002 OpenAI\u306e\u5f31\u3044NLP\u30d9\u30f3\u30c1\u30de\u30fc\u30af\u306e\u5b9f\u9a13\u3067\u306f\u3001\u7d44\u307f\u5408\u308f\u305b\u30a2\u30d7\u30ed\u30fc\u30c1\u304c\u30a2\u30e9\u30a4\u30e1\u30f3\u30c8\u3092\u6539\u5584\u3059\u308b\u3053\u3068\u304c\u793a\u3055\u308c\u3066\u3044\u308b\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2501.13124v1\">\u8ad6\u6587<\/a>\u00a0\u00a0<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2501.13124v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>\u00a0 \u00a0(Tue, 21 Jan 2025 05:36:13 GMT)<\/li>\n\n\n\n<li>\u300cSpecifically, we investigate ways of improving human supervision with a strong pretrained model and then supervise the strong model with enhanced weak human supervision.\u300d\u3068\u3044\u3046\u8a71\u3067\u8208\u5473\u6df1\u3044\u304c\u3001\u300chumans will only be able to weakly supervise superhuman models\u300d\u3068\u3044\u3046\u5371\u6a5f\u611f\u304c\u6642\u671f\u5c1a\u65e9\u3067\u3082\u306a\u3055\u305d\u3046\u306a\u306e\u304c\u30fb\u30fb\u30fb\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":[443],"class_list":["post-6132","post","type-post","status-publish","format-standard","hentry","category-arxiv","tag-weak-to-strong"],"_links":{"self":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/6132","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=6132"}],"version-history":[{"count":0,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/6132\/revisions"}],"wp:attachment":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=6132"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=6132"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=6132"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}