
{"id":10589,"date":"2026-08-01T00:33:31","date_gmt":"2026-07-31T16:33:31","guid":{"rendered":"https:\/\/infernews.com\/blog\/skillrise\/"},"modified":"2026-08-01T00:33:31","modified_gmt":"2026-07-31T16:33:31","slug":"skillrise","status":"publish","type":"post","link":"https:\/\/infernews.com\/blog\/skillrise\/","title":{"rendered":"SkillRise \u628a\u6280\u80fd\u6587\u4ef6\u8b8a\u6210\u53ef\u7d2f\u7a4d\u5b78\u7fd2"},"content":{"rendered":"\n<figure class=\"wp-block-image size-large\"><img decoding=\"async\" src=\"https:\/\/infernews.com\/blog\/wp-content\/uploads\/2026\/08\/method.jpg\" alt=\"SkillRise method\"><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">\u505a\u5b8c\u4e00\u984c\u5c31\u628a\u7d93\u9a57\u4e1f\u6389\uff0c\u5f80\u5f80\u662f\u4ee3\u7406\u7cfb\u7d71\u8a13\u7df4\u6700\u53ef\u60dc\u7684\u5730\u65b9\u3002SkillRise \u5c6c\u65bc<strong>\u5f37\u5316\u5b78\u7fd2\u6846\u67b6<\/strong>\uff0c\u7126\u9ede\u653e\u5728 cross-task skill learning\uff1a\u8b93\u540c\u4e00\u500b policy \u6309\u6b21\u5e8f\u8655\u7406\u540c\u4e00\u5bb6\u65cf\u7684\u4efb\u52d9\uff0c\u4e00\u908a\u89e3\u984c\uff0c\u4e00\u908a\u628a\u8ecc\u8de1\u6574\u7406\u6210\u6703\u6301\u7e8c\u6f14\u5316\u7684 skill document\uff0c\u5c07\u524d\u4e00\u984c\u5b78\u5230\u7684\u505a\u6cd5\u5e36\u53bb\u4e0b\u4e00\u984c\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u5b83\u7684\u53d6\u5411\u4e0d\u662f\u628a\u6bcf\u500b\u4efb\u52d9\u5206\u958b\u8a13\u7df4\u5230\u6700\u597d\uff0c\u800c\u662f\u523b\u610f\u5b89\u6392\u7531\u6dfa\u5165\u6df1\u7684\u4efb\u52d9\u5e8f\u5217\uff0c\u8b93 Solve \u8207 Curate \u4ea4\u66ff\u767c\u751f\u3002\u9019\u500b\u8a2d\u8a08\u91dd\u5c0d\u7684\u662f\u8de8\u4efb\u52d9\u9077\u79fb\u80fd\u529b\uff0c\u800c\u4e0d\u662f\u55ae\u4e00\u56de\u5408\u8868\u73fe\uff1b\u4ee3\u50f9\u662f\u74b0\u5883\u8a2d\u5b9a\u8207\u8cc7\u6599\u7d44\u7e54\u8f03\u8b1b\u7a76\uff0cALFWorld\u3001WebShop \u8981\u8ddf\u96a8 verl-agent \u7684\u74b0\u5883\u914d\u7f6e\uff0cScienceWorld \u5247\u6cbf\u7528 BEACON \u7684 setup\uff0c\u4e26\u4e14\u8981\u5148\u6574\u7406\u6a21\u578b\u8def\u5f91\u3001\u8cc7\u6599\u8def\u5f91\u8207\u8ffd\u8e64\u8a2d\u5b9a\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">README \u63d0\u4f9b\u4e86\u53ef\u76f4\u63a5\u5c0d\u7167\u7684\u57f7\u884c\u65b9\u5f0f\uff1a\u540c\u4e00\u5957 examples \u7d50\u69cb\u4e0b\uff0c\u65e2\u6709 SkillRise\uff0c\u4e5f\u6709 GRPO baseline\uff0c\u65b9\u4fbf\u628a\u65b0\u65b9\u6cd5\u8207\u57fa\u7dda\u653e\u5728\u76f8\u8fd1\u689d\u4ef6\u4e0b\u6bd4\u8f03\u3002\u6a21\u578b\u90e8\u5206\u5f9e\u8173\u672c\u540d\u7a31\u53ef\u898b\u5df2\u6e96\u5099 Qwen3-4B \u914d\u7f6e\uff0c\u5e95\u5c64\u4e5f\u5efa\u7acb\u5728 veRL\u3001verl-agent\u3001BEACON \u7b49\u73fe\u6210\u9805\u76ee\u4e4b\u4e0a\uff0c\u6240\u4ee5\u5b83\u6bd4\u8f03\u50cf\u7814\u7a76\u8207\u5be6\u9a57\u5de5\u4f5c\u6d41\u7684\u5ef6\u4f38\uff0c\u800c\u4e0d\u662f\u5373\u88dd\u5373\u7528\u7684\u7522\u54c1\u3002<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>\u628a\u300c\u89e3\u4efb\u52d9\u300d\u8207\u300c\u6574\u7406\u6280\u80fd\u300d\u62c6\u6210\u5169\u500b\u4ea4\u66ff\u89d2\u8272<\/li>\n<li>\u7528\u540c\u4e00\u5bb6\u65cf\u3001\u9010\u6b65\u8b8a\u96e3\u7684\u4efb\u52d9\u5e8f\u5217\u6e2c\u8a66\u6280\u80fd\u8f49\u79fb<\/li>\n<li>\u5728 ALFWorld\u3001WebShop\u3001ScienceWorld \u90fd\u6709\u8a55\u4f30<\/li>\n<li>README \u660e\u78ba\u4fdd\u7559 GRPO baseline \u65b9\u4fbf\u505a\u5c0d\u7167<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">\u6210\u679c\u63cf\u8ff0\u6307\u5411\u4e00\u500b\u6e05\u695a\u7d50\u8ad6\uff1aSkillRise \u5728 ALFWorld\u3001WebShop\u3001ScienceWorld \u7684\u6574\u9ad4\u7d50\u679c\u6700\u597d\uff0c\u52dd\u904e prompting-based methods \u8207 RL baselines\u3002\u8f03\u9069\u5408\u7814\u7a76 Agentic workflow\u3001\u9577\u7a0b\u6280\u80fd\u7d2f\u7a4d\u3001\u8de8\u4efb\u52d9\u5b78\u7fd2\u7684\u5718\u968a\uff1b\u60f3\u89c0\u5bdf skill document \u5982\u4f55\u5f71\u97ff\u5f8c\u7e8c\u6c7a\u7b56\u7684\u4eba\uff0c\u4e5f\u6703\u6bd4\u53ea\u770b\u6700\u7d42\u5206\u6578\u5f97\u5230\u66f4\u591a\u8a0a\u606f\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/github.com\/Within-yao\/SkillRise\" rel=\"noopener noreferrer\"><strong>GitHub<\/strong><\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>\u540c\u4e00\u500b\u7b56\u7565\u9023\u7e8c\u89e3\u5e7e\u985e\u76f8\u95dc\u4efb\u52d9\uff0c\u908a\u505a\u908a\u6574\u7406\u53ef\u8f49\u79fb\u7684\u6280\u80fd\u6587\u4ef6\uff0c\u6b63\u662f SkillRise \u60f3\u8655\u7406\u7684\u5b78\u7fd2\u65b7\u5c64\u3002\u5b83\u4e0d\u662f\u53ea\u8ffd\u55ae\u6b21\u6210\u529f\u7387\uff0c\u800c\u662f\u60f3\u8b93\u7d93\u9a57\u5728\u4e0b\u4e00\u984c\u7e7c\u7e8c\u751f\u6548\u3002<\/p>\n","protected":false},"author":8,"featured_media":10588,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"ai_generated_summary":"","footnotes":""},"categories":[133,185,116,161,201],"tags":[],"class_list":["post-10589","post","type-post","status-publish","format-standard","hentry","category-133","category-qwen","category-agentic","category-161","category-skill-"],"_links":{"self":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts\/10589","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/users\/8"}],"replies":[{"embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/comments?post=10589"}],"version-history":[{"count":0,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts\/10589\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/media\/10588"}],"wp:attachment":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/media?parent=10589"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/categories?post=10589"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/tags?post=10589"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}