{"id":4824,"date":"2024-05-06T04:44:00","date_gmt":"2024-05-05T19:44:00","guid":{"rendered":"https:\/\/devneko.jp\/wordpress\/?p=4824"},"modified":"2024-05-06T04:44:00","modified_gmt":"2024-05-05T19:44:00","slug":"better-faster-large-language-models-via-multi-token-prediction","status":"publish","type":"post","link":"https:\/\/devneko.jp\/wordpress\/?p=4824","title":{"rendered":"Better &amp; Faster Large Language Models via Multi-token Prediction"},"content":{"rendered":"\n<ul class=\"wp-block-list\">\n<li><strong>Better &amp; Faster Large Language Models via Multi-token Prediction\u00a0<\/strong>[29.1]<br>GPT\u3084Llama\u306e\u3088\u3046\u306a\u5927\u898f\u6a21\u8a00\u8a9e\u30e2\u30c7\u30eb\u306f\u3001\u6b21\u306e\u30c8\u30fc\u30b1\u30f3\u4e88\u6e2c\u640d\u5931\u3067\u8a13\u7df4\u3055\u308c\u308b\u3002 \u8907\u6570\u306e\u672a\u6765\u306e\u30c8\u30fc\u30af\u30f3\u3092\u540c\u6642\u306b\u4e88\u6e2c\u3059\u308b\u305f\u3081\u306e\u8a00\u8a9e\u30e2\u30c7\u30eb\u3092\u30c8\u30ec\u30fc\u30cb\u30f3\u30b0\u3059\u308b\u3053\u3068\u3067\u3001\u3088\u308a\u9ad8\u3044\u30b5\u30f3\u30d7\u30eb\u52b9\u7387\u304c\u5f97\u3089\u308c\u308b\u3053\u3068\u3092\u63d0\u6848\u3059\u308b\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2404.19737v1\">\u8ad6\u6587<\/a>\u00a0\u00a0<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2404.19737v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>\u00a0 \u00a0(Tue, 30 Apr 2024 17:33:57 GMT)<\/li>\n\n\n\n<li>\u6b63\u76f4\u30a2\u30a4\u30c7\u30a2\u3068\u3057\u3066\u306f\u3088\u304f\u805e\u304f\u4e88\u6e2c\u5bfe\u8c61\u306e\u8907\u7dda\u5316\u3001\u300cOur experiments (up to 7B parameters and 1T tokens) show that this is increasingly useful for larger models and in particular show strong improvements for code tasks.\u300d\u3068\u306e\u3053\u3068\u3002\u5b9f\u9a13\u7684\u306b\u793a\u3057\u305f\u306e\u306f\u91cd\u8981\u306a\u6210\u679c\u3067\u3042\u308b\u3068\u601d\u3046\u3002<\/li>\n\n\n\n<li>\u7d50\u679c\u306e\u89e3\u91c8\u3082\u53c2\u8003\u306b\u306a\u308b\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":[223],"class_list":["post-4824","post","type-post","status-publish","format-standard","hentry","category-arxiv","tag-llm"],"_links":{"self":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/4824","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=4824"}],"version-history":[{"count":0,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/4824\/revisions"}],"wp:attachment":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=4824"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=4824"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=4824"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}