{"id":3379,"date":"2023-06-05T04:30:00","date_gmt":"2023-06-04T19:30:00","guid":{"rendered":"https:\/\/devneko.jp\/wordpress\/?p=3379"},"modified":"2023-06-05T04:30:00","modified_gmt":"2023-06-04T19:30:00","slug":"bloom1-adding-language-support-to-bloom-for-zero-shot-prompting","status":"publish","type":"post","link":"https:\/\/devneko.jp\/wordpress\/?p=3379","title":{"rendered":"BLOOM+1: Adding Language Support to BLOOM for Zero-Shot Prompting\u00a0"},"content":{"rendered":"\n<ul class=\"wp-block-list\">\n<li><strong>BLOOM+1: Adding Language Support to BLOOM for Zero-Shot Prompting&nbsp;<\/strong>[50.2]<br>BLOOM\u30e2\u30c7\u30eb\u306f\u5e83\u304f\u516c\u958b\u3055\u308c\u3066\u3044\u308b\u591a\u8a00\u8a9e\u8a00\u8a9e\u30e2\u30c7\u30eb\u3067\u3042\u308b\u304c\u3001\u4e8b\u524d\u8a13\u7df4\u306f46\u8a00\u8a9e\u306b\u9650\u3089\u308c\u3066\u3044\u305f\u3002 \u65e2\u5b58\u306e\u8a00\u8a9e\u9069\u5fdc\u6226\u7565\u3092BLOOM\u306b\u9069\u7528\u3057\u30018\u3064\u306e\u65b0\u3057\u3044\u8a00\u8a9e\u306e\u6027\u80fd\u5411\u4e0a\u3092\u4fc3\u3059\u30bc\u30ed\u30b7\u30e7\u30c3\u30c8\u3092\u30d9\u30f3\u30c1\u30de\u30fc\u30af\u3059\u308b\u3002 \u30c7\u30fc\u30bf\u8a00\u8a9e\u3092\u5341\u5206\u306b\u8a13\u7df4\u3059\u308c\u3070\u3001\u591a\u69d8\u306a\u8a00\u8a9e\u306b\u9069\u5fdc\u3067\u304d\u308b\u3068\u7d50\u8ad6\u4ed8\u3051\u3066\u3044\u308b\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2212.09535v2\">\u8ad6\u6587<\/a>&nbsp;&nbsp;<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2212.09535v2\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>&nbsp; &nbsp;(Thu, 25 May 2023 10:50:40 GMT)<\/li>\n\n\n\n<li>BLOOM\u306b\u65b0\u305f\u306a\u8a00\u8a9e\u3092\u5165\u308c\u308b\u5834\u5408\u306b\u6709\u52b9\u306a\u65b9\u6cd5\u306b\u95a2\u3059\u308b\u5831\u544a\u3002\u65e5\u672c\u8a9e\u3092\u5165\u308c\u305f\u3044\u306e\u3067\u3068\u3066\u3082\u6709\u7528\u306a\u60c5\u5831\u3002<\/li>\n\n\n\n<li>\u300cSurprisingly, we find that adapter-based finetuning is more effective than continued pretraining for large models.\u300d\u3068\u3044\u3046\u9762\u767d\u3044\u7d50\u679c\u304c\u5831\u544a\u3055\u308c\u3066\u3044\u308b\u3002\u300cwe need around 100 million tokens of the new language for effective language adaptation.\u300d\u3068\u3044\u3046\u306e\u3082\u9762\u767d\u3044\u3002wikipedia\u30ec\u30d9\u30eb\u3067\u5341\u5206\u3068\u3044\u3046\u306e\u306f\u672c\u5f53\u306a\u3093\u3060\u308d\u3046\u304b\u3002<\/li>\n\n\n\n<li>\u30ea\u30dd\u30b8\u30c8\u30ea\u306f<a href=\"https:\/\/github.com\/bigscience-workshop\/multilingual-modeling\">GitHub &#8211; bigscience-workshop\/multilingual-modeling: Adapting BLOOM model to support a new unseen language<\/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":[223,267],"class_list":["post-3379","post","type-post","status-publish","format-standard","hentry","category-arxiv","tag-llm","tag-multilingual"],"_links":{"self":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/3379","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=3379"}],"version-history":[{"count":0,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/3379\/revisions"}],"wp:attachment":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=3379"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=3379"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=3379"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}