{"id":5589,"date":"2024-10-15T05:48:00","date_gmt":"2024-10-14T20:48:00","guid":{"rendered":"https:\/\/devneko.jp\/wordpress\/?p=5589"},"modified":"2024-10-15T05:48:00","modified_gmt":"2024-10-14T20:48:00","slug":"extracting-and-transferring-abilities-for-building-multi-lingual-ability-enhanced-large-language-models","status":"publish","type":"post","link":"https:\/\/devneko.jp\/wordpress\/?p=5589","title":{"rendered":"Extracting and Transferring Abilities For Building Multi-lingual Ability-enhanced Large Language Models"},"content":{"rendered":"\n<ul class=\"wp-block-list\">\n<li><strong>Extracting and Transferring Abilities For Building Multi-lingual Ability-enhanced Large Language Models\u00a0<\/strong>[105.0]<br>\u6211\u3005\u306f,MAET\u3068\u547d\u540d\u3055\u308c\u305f\u591a\u8a00\u8a9e\u80fd\u529b\u62bd\u51fa\u3068\u4f1d\u9054\u624b\u6cd5\u3092\u63d0\u6848\u3059\u308b\u3002 \u6211\u3005\u306e\u30ad\u30fc\u3068\u306a\u308b\u8003\u3048\u65b9\u306f\u3001\u5927\u304d\u306a\u8a00\u8a9e\u30e2\u30c7\u30eb\u304b\u3089\u8a00\u8a9e\u306b\u4f9d\u5b58\u3057\u306a\u3044\u80fd\u529b\u306b\u95a2\u3059\u308b\u91cd\u307f\u3092\u5206\u89e3\u3057\u62bd\u51fa\u3059\u308b\u3053\u3068\u3067\u3042\u308b\u3002 \u5b9f\u9a13\u7d50\u679c\u304b\u3089,MAET\u306f\u9ad8\u5ea6\u80fd\u529b\u306e\u62bd\u51fa\u3068\u4f1d\u9054\u3092\u52b9\u679c\u7684\u306b\u884c\u3046\u3053\u3068\u304c\u3067\u304d,\u30c8\u30ec\u30fc\u30cb\u30f3\u30b0\u30d9\u30fc\u30b9\u30e9\u30a4\u30f3\u6cd5\u3088\u308a\u3082\u512a\u308c\u308b\u3053\u3068\u304c\u308f\u304b\u3063\u305f\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2410.07825v1\">\u8ad6\u6587<\/a>\u00a0\u00a0<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2410.07825v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>\u00a0 \u00a0(Thu, 10 Oct 2024 11:23:18 GMT)<\/li>\n\n\n\n<li>\u300cOur key idea is to decompose and extract language-agnostic ability-related weights from LLMs, and transfer them across different languages by simple addition and subtraction operations without training.\u300d\u3068\u3044\u3046\u591a\u8a00\u8a9e\u80fd\u529b\u306e\u62bd\u51fa\u3068\u305d\u306e\u30e2\u30c7\u30eb\u30de\u30fc\u30b8\u624b\u6cd5\u3001MEAT: Multi-lingual Ability Extraction and Transfer approach\u3092\u63d0\u6848\u3002\u300cOur approach MAET achieves better performance than the competitive baseline methods (e g , continual pre-training and model merging with task vector) in multi-lingual complex reasoning tasks, including mathematical reasoning tasks and scientific reasoning tasks.\u300d\u3068\u306e\u3053\u3068\u3002<\/li>\n\n\n\n<li>\u30ea\u30dd\u30b8\u30c8\u30ea\u306f<a href=\"https:\/\/github.com\/RUCAIBox\/MAET\">https:\/\/github.com\/RUCAIBox\/MAET<\/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":[256,267],"class_list":["post-5589","post","type-post","status-publish","format-standard","hentry","category-arxiv","tag-model-merging","tag-multilingual"],"_links":{"self":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/5589","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=5589"}],"version-history":[{"count":0,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/5589\/revisions"}],"wp:attachment":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=5589"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=5589"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=5589"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}