{"id":5426,"date":"2024-09-12T05:10:00","date_gmt":"2024-09-11T20:10:00","guid":{"rendered":"https:\/\/devneko.jp\/wordpress\/?p=5426"},"modified":"2024-09-12T05:10:00","modified_gmt":"2024-09-11T20:10:00","slug":"the-representation-landscape-of-few-shot-learning-and-fine-tuning-in-large-language-models","status":"publish","type":"post","link":"https:\/\/devneko.jp\/wordpress\/?p=5426","title":{"rendered":"The representation landscape of few-shot learning and fine-tuning in large language models"},"content":{"rendered":"\n<ul class=\"wp-block-list\">\n<li><strong>The representation landscape of few-shot learning and fine-tuning in large language models\u00a0<\/strong>[43.8]<br>In-context Learning (ICL) \u3068 supervised Fine-tuning (SFT) \u306f\u3001\u73fe\u4ee3\u306e\u5927\u898f\u6a21\u8a00\u8a9e\u30e2\u30c7\u30eb (LLM) \u306e\u6027\u80fd\u5411\u4e0a\u306e\u305f\u3081\u306e2\u3064\u306e\u4e00\u822c\u7684\u306a\u6226\u7565\u3067\u3042\u308b\u3002 \u3053\u306e2\u3064\u306e\u4e8b\u4f8b\u306b\u304a\u3044\u3066,\u96a0\u308c\u305f\u8868\u73fe\u306e\u78ba\u7387\u7684\u666f\u89b3\u3092\u89e3\u6790\u3057\u305f\u3002 ICL\u3068SFT\u306f\u3001\u3069\u3061\u3089\u3082\u30cd\u30c3\u30c8\u30ef\u30fc\u30af\u306e\u4e2d\u592e\u3067\u6025\u6fc0\u306a\u9077\u79fb\u3092\u884c\u3046\u5834\u5408\u306b\u304a\u3044\u3066\u3001\u975e\u5e38\u306b\u7570\u306a\u308b\u5185\u90e8\u69cb\u9020\u3092\u751f\u6210\u3059\u308b\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2409.03662v1\">\u8ad6\u6587<\/a>\u00a0\u00a0<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2409.03662v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>\u00a0 \u00a0(Thu, 5 Sep 2024 16:15:12 GMT)<\/li>\n\n\n\n<li>ICL\u3068SFT\u306e\u52d5\u4f5c\u306e\u5dee\u306e\u5206\u6790\u3001\u300cwe compare how LLMs solve the same question-answering task, finding that ICL and SFT create very different internal structures, in both cases undergoing a sharp transition in the middle of the network.\u300d\u3068\u306e\u3053\u3068\u3067\u6319\u52d5\u304c\u304b\u306a\u308a\u7570\u306a\u308b\u3088\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":[150,200],"class_list":["post-5426","post","type-post","status-publish","format-standard","hentry","category-arxiv","tag-fine-tuning","tag-in-context-learning"],"_links":{"self":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/5426","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=5426"}],"version-history":[{"count":0,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/5426\/revisions"}],"wp:attachment":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=5426"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=5426"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=5426"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}