{"id":5114,"date":"2024-07-08T05:47:00","date_gmt":"2024-07-07T20:47:00","guid":{"rendered":"https:\/\/devneko.jp\/wordpress\/?p=5114"},"modified":"2024-07-08T05:47:00","modified_gmt":"2024-07-07T20:47:00","slug":"mixture-of-insightful-experts-mote-the-synergy-of-thought-chains-and-expert-mixtures-in-self-alignment","status":"publish","type":"post","link":"https:\/\/devneko.jp\/wordpress\/?p=5114","title":{"rendered":"Mixture of insighTful Experts (MoTE): The Synergy of Thought Chains and Expert Mixtures in Self-Alignment"},"content":{"rendered":"\n<ul class=\"wp-block-list\">\n<li><strong>Mixture of insighTful Experts (MoTE): The Synergy of Thought Chains and Expert Mixtures in Self-Alignment\u00a0<\/strong>[103.1]<br>\u5f93\u6765\u306e\u30a2\u30e9\u30a4\u30e1\u30f3\u30c8\u6226\u7565\u306f\u4eba\u9593\u306e\u4ecb\u5165\u306b\u5927\u304d\u304f\u4f9d\u5b58\u3057\u3066\u304a\u308a\u3001\u4f8b\u3048\u3070Supervised Fine-Tuning(SFT)\u3084Reinforcement Learning from Human Feedback(RLHF)\u306a\u3069\u3067\u3042\u308b\u3002 \u672c\u7a3f\u3067\u306f\u3001AlignCoT\u3068\u547c\u3070\u308c\u308b\u601d\u8003\u306e\u9023\u9396(CoT)\u30a2\u30d7\u30ed\u30fc\u30c1\u3092\u5229\u7528\u3057\u305f\u65b0\u3057\u3044\u81ea\u5df1\u30a2\u30e9\u30a4\u30e1\u30f3\u30c8\u624b\u6cd5\u3092\u63d0\u6848\u3059\u308b\u3002 \u672c\u7a3f\u3067\u306f\u3001AlignCoT\u30d7\u30ed\u30bb\u30b9\u306e\u5404\u30b3\u30f3\u30dd\u30fc\u30cd\u30f3\u30c8\u3092\u5f37\u5316\u3059\u308b\u305f\u3081\u306b\u5c02\u9580\u5bb6\u306e\u6df7\u5408\u3092\u9069\u7528\u3057\u3001\u30a2\u30e9\u30a4\u30e1\u30f3\u30c8\u52b9\u7387\u3092\u8457\u3057\u304f\u5411\u4e0a\u3055\u305b\u308bMoTE\u30a2\u30fc\u30ad\u30c6\u30af\u30c1\u30e3\u306b\u3064\u3044\u3066\u7d39\u4ecb\u3059\u308b\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2405.00557v2\">\u8ad6\u6587<\/a>\u00a0\u00a0<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2405.00557v2\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>\u00a0 \u00a0(Wed, 03 Jul 2024 15:04:25 GMT)<\/li>\n\n\n\n<li>CoT\u7684\u624b\u6cd5\u3092\u7528\u3044\u305f\u81ea\u5df1\u30a2\u30e9\u30a4\u30e1\u30f3\u30c8\u624b\u6cd5ALignCoT \u3068\u3055\u3089\u306b\u305d\u308c\u3092\u52b9\u7387\u5316\u3059\u308bMoTE\uff08Mixture of insighTful Experts\uff09\u306e\u63d0\u6848\u3002<\/li>\n\n\n\n<li>\u300cSafety alignment is essential for LLMs.Existing approaches like SFT and RLHF rely extensively on human annotation, whereas self-alignment strategies depend on LLMs\u2019 emergent abilities.\u300d\u306f\u305d\u3046\u306a\u3093\u3060\u308d\u3046\u3068\u601d\u3046\u306e\u3060\u3051\u3069\u3001\u5f37\u529b\u306a\u81ea\u5df1\u30a2\u30e9\u30a4\u30e1\u30f3\u30c8\u624b\u6cd5\u304c\u5b89\u5168\u306b\u3064\u306a\u304c\u308b\u306e\u304b\u306f\u3069\u3046\u306a\u3093\u3060\u308d\u3046\u3002\u3002\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":[22,59,223],"class_list":["post-5114","post","type-post","status-publish","format-standard","hentry","category-arxiv","tag-alignment","tag-chain-of-thought","tag-llm"],"_links":{"self":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/5114","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=5114"}],"version-history":[{"count":0,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/5114\/revisions"}],"wp:attachment":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=5114"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=5114"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=5114"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}