{"id":4822,"date":"2024-05-09T04:39:00","date_gmt":"2024-05-08T19:39:00","guid":{"rendered":"https:\/\/devneko.jp\/wordpress\/?p=4822"},"modified":"2024-05-09T04:39:00","modified_gmt":"2024-05-08T19:39:00","slug":"is-bigger-edit-batch-size-always-better-an-empirical-study-on-model-editing-with-llama-3","status":"publish","type":"post","link":"https:\/\/devneko.jp\/wordpress\/?p=4822","title":{"rendered":"Is Bigger Edit Batch Size Always Better? &#8212; An Empirical Study on Model Editing with Llama-3\u00a0"},"content":{"rendered":"\n<ul class=\"wp-block-list\">\n<li><strong>Is Bigger Edit Batch Size Always Better? &#8212; An Empirical Study on Model Editing with Llama-3\u00a0<\/strong>[2.6]<br>\u672c\u7814\u7a76\u3067\u306f,\u6700\u65b0\u306e\u5927\u8a00\u8a9e\u30e2\u30c7\u30eb\u3067\u3042\u308bLlama-3\u306b\u7740\u76ee\u3057\u305f\u30bf\u30fc\u30b2\u30c3\u30c8\u30e2\u30c7\u30eb\u7de8\u96c6\u5206\u6790\u3092\u884c\u3046\u3002 \u6700\u59274096\u500b\u306e\u7de8\u96c6\u3092\u5bfe\u8c61\u3068\u3059\u308b\u8a55\u4fa1\u306b\u3088\u308a,\u6700\u3082\u52b9\u679c\u7684\u306a\u7de8\u96c6\u5c64\u3092\u540c\u5b9a\u3059\u308b\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2405.00664v1\">\u8ad6\u6587<\/a>\u00a0\u00a0<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2405.00664v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>\u00a0 \u00a0(Wed, 01 May 2024 17:50:37 GMT)<\/li>\n\n\n\n<li>Llama-3\u3092\u5bfe\u8c61\u3068\u3057\u305f\u30e2\u30c7\u30eb\u7de8\u96c6\u3001\u51fa\u308b\u306e\u304c\u901f\u3044\u30fb\u30fb\u30fb<\/li>\n\n\n\n<li>\u300cContrary to previous belief, our experiments show that earlier layers may be more optimal intervention points, and that smaller, frequent sequential batch size edits have a superior performance in comparison to larger batch sizes.\u300d\u3001\u3053\u306e\u624b\u306e\u30c6\u30af\u30cb\u30c3\u30af\u306f\u30e2\u30c7\u30eb\u304c\u66f4\u65b0\u3055\u308c\u308b\u305f\u3073\u5909\u308f\u308b\u306e\u3060\u308d\u3046\u304b\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":[222,254],"class_list":["post-4822","post","type-post","status-publish","format-standard","hentry","category-arxiv","tag-llama-3","tag-model-editing"],"_links":{"self":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/4822","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=4822"}],"version-history":[{"count":0,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/4822\/revisions"}],"wp:attachment":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=4822"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=4822"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=4822"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}