{"id":6395,"date":"2025-03-10T04:03:00","date_gmt":"2025-03-09T19:03:00","guid":{"rendered":"https:\/\/devneko.jp\/wordpress\/?p=6395"},"modified":"2025-03-10T04:03:00","modified_gmt":"2025-03-09T19:03:00","slug":"implicit-cross-lingual-rewarding-for-efficient-multilingual-preference-alignment","status":"publish","type":"post","link":"https:\/\/devneko.jp\/wordpress\/?p=6395","title":{"rendered":"Implicit Cross-Lingual Rewarding for Efficient Multilingual Preference Alignment\u00a0"},"content":{"rendered":"\n<ul class=\"wp-block-list\">\n<li><strong>Implicit Cross-Lingual Rewarding for Efficient Multilingual Preference Alignment\u00a0<\/strong>[35.2]<br>\u63d0\u6848\u624b\u6cd5\u306f,\u6697\u9ed9\u306e\u5831\u5968\u306b\u3088\u3063\u3066\u9069\u5207\u306b\u6574\u5408\u3057\u305f\u82f1\u8a9e\u30e2\u30c7\u30eb\u304b\u3089\u306e\u597d\u307f\u3092\u6349\u3048,\u53cd\u5fa9\u5b66\u7fd2\u3092\u901a\u3058\u3066\u4ed6\u8a00\u8a9e\u306b\u4f1d\u9054\u3059\u308b\u624b\u6cd5\u3067\u3042\u308b\u3002 2\u56de\u306b\u5206\u3051\u3066\u5fae\u8abf\u6574\u3057\u305fLlama3\u306f\u30a6\u30a3\u30f3\u30ec\u30fc\u30c8\u3092\u5e73\u574712.72%\u6539\u5584\u3057\u3001X-AlpacaEval\u306e\u30ea\u30fc\u30c0\u30fc\u30dc\u30fc\u30c9\u4e0a\u3067\u306e\u30c8\u30ec\u30fc\u30cb\u30f3\u30b0\u8a00\u8a9e\u5168\u4f53\u306e\u9577\u3055\u5236\u5fa1\u30a6\u30a3\u30f3\u30ec\u30fc\u30c8\u30925.97%\u5411\u4e0a\u3055\u305b\u305f\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2503.04647v1\">\u8ad6\u6587<\/a>\u00a0\u00a0<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2503.04647v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>\u00a0 \u00a0(Thu, 06 Mar 2025 17:33:01 GMT)<\/li>\n\n\n\n<li>\u300cwe propose a novel approach that captures learned preferences from well-aligned English models by implicit rewards and transfers them to other languages through iterative training.\u300d\u3001\u3068\u306e\u3053\u3068\u3067\u82f1\u8a9e\u306e\u9078\u597d\u3092\u30de\u30eb\u30c1\u30ea\u30f3\u30ac\u30eb\u306b\u8ee2\u9001\u3059\u308b\u624b\u6cd5\u306e\u63d0\u6848\u3002\u300cMultilingual Responses Generation\u3001Implicit Cross-lingual Rewarding\u3001Preference Transfer Training\u300d\u306e\uff13\u3064\u304b\u3089\u306a\u308b<\/li>\n\n\n\n<li>\u30ea\u30dd\u30b8\u30c8\u30ea\u306f<a href=\"https:\/\/github.com\/ZNLP\/Implicit-Cross-Lingual-Rewarding\">GitHub &#8211; ZNLP\/Implicit-Cross-Lingual-Rewarding<\/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":[267],"class_list":["post-6395","post","type-post","status-publish","format-standard","hentry","category-arxiv","tag-multilingual"],"_links":{"self":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/6395","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=6395"}],"version-history":[{"count":0,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/6395\/revisions"}],"wp:attachment":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=6395"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=6395"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=6395"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}