{"id":4563,"date":"2024-03-18T06:55:00","date_gmt":"2024-03-17T21:55:00","guid":{"rendered":"https:\/\/devneko.jp\/wordpress\/?p=4563"},"modified":"2024-03-18T06:55:00","modified_gmt":"2024-03-17T21:55:00","slug":"easy-to-hard-generalization-scalable-alignment-beyond-human-supervision","status":"publish","type":"post","link":"https:\/\/devneko.jp\/wordpress\/?p=4563","title":{"rendered":"Easy-to-Hard Generalization: Scalable Alignment Beyond Human Supervision"},"content":{"rendered":"\n<ul class=\"wp-block-list\">\n<li><strong>Easy-to-Hard Generalization: Scalable Alignment Beyond Human Supervision&nbsp;<\/strong>[99.0]<br>\u73fe\u5728\u306eAI\u30a2\u30e9\u30a4\u30e1\u30f3\u30c8\u624b\u6cd5\u306f\u3001\u4eba\u9593\u304c\u63d0\u4f9b\u3059\u308b\u5b9f\u6f14\u3084\u5224\u65ad\u306b\u4f9d\u5b58\u3057\u3066\u3044\u308b\u3002 \u5f7c\u3089\u306e\u80fd\u529b\u304c\u4eba\u9593\u306e\u30ec\u30d9\u30eb\u3092\u8d85\u3048\u305f\u3068\u304d\u3001\u30b7\u30b9\u30c6\u30e0\u3092\u6539\u5584\u3059\u308b\u306b\u306f\u3069\u3046\u3059\u308c\u3070\u3088\u3044\u306e\u304b?<br><a href=\"http:\/\/arxiv.org\/abs\/2403.09472v1\">\u8ad6\u6587<\/a>&nbsp;&nbsp;<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2403.09472v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>&nbsp; &nbsp;(Thu, 14 Mar 2024 15:12:38 GMT)<\/li>\n\n\n\n<li><a href=\"https:\/\/devneko.jp\/wordpress\/?p=4320\">The Unreasonable Effectiveness of Easy Training Data for Hard Tasks \u2013 arXiv\u6700\u65b0\u8ad6\u6587\u306e\u7d39\u4ecb (devneko.jp)<\/a>\u3067\u3082\u53d6\u308a\u4e0a\u3052\u3089\u308c\u3066\u3044\u305f\u8a71\u3060\u304c\u3001PRMs(process reward models)\u3084OPRMs(Outcome &amp; Process Reward Model)\u3092\u7528\u3044\u308b\u3068\u3055\u3089\u306b\u6709\u52b9\u3068\u306e\u3053\u3068\u3002<\/li>\n\n\n\n<li>AGI\u3084ASI\u3068\u3044\u3046\u8a71\u3092\u805e\u304f\u306b\u3053\u306e\u3088\u3046\u306a\u624b\u6cd5\u306e\u91cd\u8981\u6027\u304c\u9ad8\u307e\u3063\u3066\u3044\u308b\u3088\u3046\u306b\u601d\u3046\uff08\u4e00\u65b9\u3067\u7d50\u8ad6\u306b\u3042\u308b\u300cThis approach presents a promising direction for developing AI systems capable of surpassing human problem-solving capabilities\u300d\u306e\u3088\u3046\u306b\u4eba\u9593\u304cEasy\u5074\u306b\u4f4d\u7f6e\u3065\u3051\u3089\u308c\u308b\u306e\u306f\u8907\u96d1\u306a\u601d\u3044\u3082\u3042\u308b\uff09<\/li>\n\n\n\n<li>\u30ea\u30dd\u30b8\u30c8\u30ea\u306f<a href=\"https:\/\/github.com\/Edward-Sun\/easy-to-hard\">Edward-Sun\/easy-to-hard (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":[22,121,576],"class_list":["post-4563","post","type-post","status-publish","format-standard","hentry","category-arxiv","tag-alignment","tag-easy2hard","tag-576"],"_links":{"self":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/4563","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=4563"}],"version-history":[{"count":0,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/4563\/revisions"}],"wp:attachment":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=4563"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=4563"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=4563"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}