{"id":5339,"date":"2024-08-29T03:38:00","date_gmt":"2024-08-28T18:38:00","guid":{"rendered":"https:\/\/devneko.jp\/wordpress\/?p=5339"},"modified":"2024-08-29T03:38:00","modified_gmt":"2024-08-28T18:38:00","slug":"performance-law-of-large-language-models","status":"publish","type":"post","link":"https:\/\/devneko.jp\/wordpress\/?p=5339","title":{"rendered":"Performance Law of Large Language Models"},"content":{"rendered":"\n<ul class=\"wp-block-list\">\n<li><strong>Performance Law of Large Language Models&nbsp;<\/strong>[58.3]<br>\u6027\u80fd\u6cd5\u5247\u306f\u3001LLM\u30a2\u30fc\u30ad\u30c6\u30af\u30c1\u30e3\u306e\u9078\u629e\u3068\u8a08\u7b97\u8cc7\u6e90\u306e\u52b9\u7387\u7684\u306a\u5272\u308a\u5f53\u3066\u3092\u5c0e\u304f\u305f\u3081\u306b\u7528\u3044\u3089\u308c\u308b\u3002 \u6027\u80fd\u6cd5\u5247\u306f\u3001LLM\u30a2\u30fc\u30ad\u30c6\u30af\u30c1\u30e3\u306e\u9078\u629e\u3068\u8a08\u7b97\u8cc7\u6e90\u306e\u52b9\u7387\u7684\u306a\u5272\u308a\u5f53\u3066\u3092\u5e83\u7bc4\u306a\u5b9f\u9a13\u306a\u3057\u3067\u5c0e\u304f\u306e\u306b\u5229\u7528\u3067\u304d\u308b\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2408.09895v1\">\u8ad6\u6587<\/a>&nbsp;&nbsp;<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2408.09895v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>&nbsp; &nbsp;(Mon, 19 Aug 2024 11:09:12 GMT)<\/li>\n\n\n\n<li>\u8a08\u7b97\u5f0f\u3067MMLU\u30b9\u30b3\u30a2\u3092\u76f4\u63a5\u4e88\u6e2c\u3001\u30ad\u30fc\u3068\u306a\u308b\u306e\u306f\u300c\u2022 The number of layers N \u2022 The hidden size h \u2022 The intermediate size d of FFN \u2022 The size of training data T (trillion tokens) \u2022 The model size S (billion parameters)\u300d\u3068\u306e\u3053\u3068<\/li>\n\n\n\n<li>\u9762\u767d\u3044\u3051\u3069\u307b\u3093\u307e\u304b\u3044\u306a<\/li>\n<\/ul>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>Large Language Monkeys: Scaling Inference Compute with Repeated Sampling\u00a0<\/strong>[81.3]<br>\u751f\u6210\u3057\u305f\u30b5\u30f3\u30d7\u30eb\u306e\u6570\u3092\u5897\u3084\u3059\u3053\u3068\u3067\u3001\u5225\u306e\u8ef8\u3068\u3057\u3066\u63a8\u8ad6\u8a08\u7b97\u3092\u63a2\u7d22\u3059\u308b\u3002 \u3059\u3079\u3066\u306e\u56de\u7b54\u3092\u81ea\u52d5\u7684\u306b\u691c\u8a3c\u3067\u304d\u308b\u30b3\u30fc\u30c7\u30a3\u30f3\u30b0\u3084\u5f62\u5f0f\u8a3c\u660e\u306e\u3088\u3046\u306a\u30c9\u30e1\u30a4\u30f3\u3067\u306f\u3001\u30ab\u30d0\u30ec\u30c3\u30b8\u306e\u5897\u52a0\u306f\u76f4\u63a5\u7684\u306b\u30d1\u30d5\u30a9\u30fc\u30de\u30f3\u30b9\u306e\u5411\u4e0a\u306b\u5909\u63db\u3055\u308c\u308b\u3002 \u591a\u304f\u306e\u4e16\u4ee3\u304b\u3089\u6b63\u3057\u3044\u30b5\u30f3\u30d7\u30eb\u3092\u540c\u5b9a\u3059\u308b\u3053\u3068\u306f\u3001\u81ea\u52d5\u691c\u8a3c\u306e\u306a\u3044\u9818\u57df\u306b\u304a\u3051\u308b\u5c06\u6765\u306e\u7814\u7a76\u306b\u3068\u3063\u3066\u91cd\u8981\u306a\u65b9\u5411\u3067\u3042\u308b\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2407.21787v1\">\u8ad6\u6587<\/a>\u00a0\u00a0<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2407.21787v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>\u00a0 \u00a0(Wed, 31 Jul 2024 17:57:25 GMT)<\/li>\n\n\n\n<li>\u8a08\u7b97\u80fd\u529b\u3068\u3044\u3046\u9762\u3067\u306f\u3001\u63a8\u8ad6\u5074\u3067\u306eScaling\u3068\u3044\u3046\u8a71\u3082<\/li>\n\n\n\n<li>\uff08\u3053\u306e\u8fba\u306f\u826f\u8cea\u306a\u5408\u6210\u30c7\u30fc\u30bf\u3068\u306e\u95a2\u4fc2\u6027\u3082\u6c17\u306b\u306a\u308b\uff09<\/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":[349],"class_list":["post-5339","post","type-post","status-publish","format-standard","hentry","category-arxiv","tag-scaling-law"],"_links":{"self":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/5339","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=5339"}],"version-history":[{"count":0,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/5339\/revisions"}],"wp:attachment":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=5339"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=5339"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=5339"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}