{"id":4441,"date":"2024-02-19T03:25:00","date_gmt":"2024-02-18T18:25:00","guid":{"rendered":"https:\/\/devneko.jp\/wordpress\/?p=4441"},"modified":"2024-02-19T03:25:00","modified_gmt":"2024-02-18T18:25:00","slug":"data-engineering-for-scaling-language-models-to-128k-context","status":"publish","type":"post","link":"https:\/\/devneko.jp\/wordpress\/?p=4441","title":{"rendered":"Data Engineering for Scaling Language Models to 128K Context"},"content":{"rendered":"\n<ul class=\"wp-block-list\">\n<li><strong>Data Engineering for Scaling Language Models to 128K Context\u00a0<\/strong>[98.4]<br>\u672c\u7814\u7a76\u3067\u306f,\u8a00\u8a9e\u30e2\u30c7\u30eb\u306e\u6587\u8108\u9577\u3092128K\u307e\u3067\u62e1\u5f35\u3059\u308b\u305f\u3081\u306e\u7d99\u7d9a\u4e8b\u524d\u5b66\u7fd2\u6cd5\u306b\u3064\u3044\u3066\u691c\u8a0e\u3059\u308b\u3002\u00a0\u9577\u3044\u30b3\u30f3\u30c6\u30ad\u30b9\u30c8\u30e2\u30c7\u30ea\u30f3\u30b0\u3001\u7279\u306bthe ability to use information at any input locations \u306f\u3001\u4e3b\u306b\u5927\u898f\u6a21\u4e8b\u524d\u30c8\u30ec\u30fc\u30cb\u30f3\u30b0\u306b\u3088\u3063\u3066\u65e2\u306b\u7372\u5f97\u3055\u308c\u3066\u3044\u308b\u6a5f\u80fd\u3067\u3042\u308a\u3001\u3053\u306e\u80fd\u529b\u306f\u3001\u9069\u5207\u306a\u30c7\u30fc\u30bf\u6df7\u5408\u4e0a\u3067\u306e\u8efd\u91cf\u306a\u9023\u7d9a\u7684\u4e8b\u524d\u30c8\u30ec\u30fc\u30cb\u30f3\u30b0\u3092\u901a\u3058\u3066\u3001\u30c8\u30ec\u30fc\u30cb\u30f3\u30b0\u4e2d(\u4f8b\u3048\u3070\u30014k\u304b\u3089128k)\u306b\u304a\u3044\u3066\u3001\u304b\u306a\u308a\u9577\u3044\u30b3\u30f3\u30c6\u30ad\u30b9\u30c8\u306b\u62e1\u5f35\u3067\u304d\u308b\u3068\u4eee\u5b9a\u3059\u308b\u3002\u00a0\u6211\u3005\u306e\u30ec\u30b7\u30d4\u306f\u5f37\u529b\u306a\u30aa\u30fc\u30d7\u30f3\u30bd\u30fc\u30b9\u9577\u6587\u30e2\u30c7\u30eb\u3088\u308a\u512a\u308c\u3066\u304a\u308a\u3001GPT-4 128K\u306e\u3088\u3046\u306a\u30d5\u30ed\u30f3\u30c6\u30a3\u30a2\u30e2\u30c7\u30eb\u3068\u306e\u30ae\u30e3\u30c3\u30d7\u3092\u57cb\u3081\u3066\u3044\u308b\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2402.10171v1\">\u8ad6\u6587<\/a>\u00a0\u00a0<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2402.10171v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>\u00a0 \u00a0(Thu, 15 Feb 2024 18:19:16 GMT)<\/li>\n\n\n\n<li>\u9577\u6587\u5bfe\u5fdc\u306e\u305f\u3081\u306e\u30ec\u30b7\u30d4\u3002\u300cthe ability to utilize information at arbitrary locations within the 128K input is already mostly acquired by large-scale pretraining, even for models pretrained on substantially shorter 4K context.\u300d\u3068\u3044\u3046\u306e\u306f\u8208\u5473\u6df1\u3044\u3002<\/li>\n\n\n\n<li>\u30ea\u30dd\u30b8\u30c8\u30ea\u306f<a href=\"https:\/\/github.com\/FranxYao\/Long-Context-Data-Engineering\">FranxYao\/Long-Context-Data-Engineering: Implementation of paper Data Engineering for Scaling Language Models to 128K Context (github.com)<\/a><\/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":[223,667],"class_list":["post-4441","post","type-post","status-publish","format-standard","hentry","category-arxiv","tag-llm","tag-667"],"_links":{"self":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/4441","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=4441"}],"version-history":[{"count":0,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/4441\/revisions"}],"wp:attachment":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=4441"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=4441"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=4441"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}