{"id":76861,"date":"2026-02-23T14:07:37","date_gmt":"2026-02-23T06:07:37","guid":{"rendered":"https:\/\/www.wsisp.com\/helps\/76861.html"},"modified":"2026-02-23T14:07:37","modified_gmt":"2026-02-23T06:07:37","slug":"2025_nips_reasonflux-prm-trajectory-aware-prms-for-long-chain-of-thought-reasoning-in-llms","status":"publish","type":"post","link":"https:\/\/www.wsisp.com\/helps\/76861.html","title":{"rendered":"2025_NIPS_ReasonFlux-PRM: Trajectory-Aware PRMs for Long Chain-of-Thought Reasoning in LLMs"},"content":{"rendered":"

\"\u5728\u8fd9\u91cc\u63d2\u5165\u56fe\u7247\u63cf\u8ff0\"<\/p>\n

\u6587\u7ae0\u6838\u5fc3\u603b\u7ed3\u4e0e\u521b\u65b0\u70b9<\/h2>\n

\u4e3b\u8981\u5185\u5bb9<\/h3>\n

\u672c\u6587\u9488\u5bf9\u73b0\u6709\u8fc7\u7a0b\u5956\u52b1\u6a21\u578b\uff08PRMs\uff09\u96be\u4ee5\u8bc4\u4f30\u5927\u8bed\u8a00\u6a21\u578b\u4e2d\u95f4\u63a8\u7406\u8f68\u8ff9\u7684\u95ee\u9898\uff0c\u63d0\u51fa\u4e86\u8f68\u8ff9\u611f\u77e5\u7684PRM\u2014\u2014ReasonFlux-PRM\u3002\u8be5\u6a21\u578b\u901a\u8fc7\u6b65\u7ea7\u548c\u8f68\u8ff9\u7ea7\u53cc\u91cd\u76d1\u7763\uff0c\u9002\u914d\u201c\u8f68\u8ff9-\u54cd\u5e94\u201d\u683c\u5f0f\u7684\u957f\u94fe\u63a8\u7406\u6570\u636e\uff0c\u53ef\u5e94\u7528\u4e8e\u79bb\u7ebf\u9ad8\u8d28\u91cf\u6570\u636e\u7b5b\u9009\u3001\u5728\u7ebf\u5f3a\u5316\u5b66\u4e60\u7b56\u7565\u4f18\u5316\u548c\u6d4b\u8bd5\u65f6Best-of-N\u7f29\u653e\u4e09\u5927\u573a\u666f\u3002\u5728AIME\u3001MATH500\u7b49\u591a\u4e2a\u6311\u6218\u6027\u57fa\u51c6\u6d4b\u8bd5\u4e2d\uff0cReasonFlux-PRM\uff087B\u7248\u672c\uff09\u8868\u73b0\u4f18\u4e8eQwen2.5-Math-PRM-72B\u7b49\u5f3a\u57fa\u7ebf\u6a21\u578b\u548c\u4eba\u5de5\u7cbe\u9009\u6570\u636e\uff0c\u5728\u76d1\u7763\u5fae\u8c03\u3001\u5f3a\u5316\u5b66\u4e60\u548c\u6d4b\u8bd5\u65f6\u7f29\u653e\u573a\u666f\u5206\u522b\u5b9e\u73b012.1%\u30014.5%\u548c6.3%\u7684\u5e73\u5747\u6027\u80fd\u63d0\u5347\uff0c\u540c\u65f6\u8fd8\u53d1\u5e03\u4e86\u9002\u7528\u4e8e\u8d44\u6e90\u53d7\u9650\u573a\u666f\u76841.5B\u8f7b\u91cf\u5316\u7248\u672c\u3002<\/p>\n

\u521b\u65b0\u70b9<\/h3>\n
  • \u8f68\u8ff9\u611f\u77e5\u5956\u52b1\u8bbe\u8ba1\uff1a\u9996\u6b21\u5c06\u6b65\u7ea7\u548c\u8f68\u8ff9\u7ea7\u76d1\u7763\u7ed3\u5408\uff0c\u6b65\u7ea7\u5956\u52b1\u5305\u542b\u8bed\u4e49\u5bf9\u9f50\u3001\u903b\u8f91\u8d28\u91cf\u548c\u4e0a\u4e0b\u6587\u8fde\u8d2f\u6027\u4e09\u4e2a\u7ef4\u5ea6\uff0c\u8f68\u8ff9\u7ea7\u5956\u52b1\u901a\u8fc7\u6a21\u677f\u5f15\u5bfc\u8bc4\u4f30\u63a8\u7406\u7b56\u7565\u7684\u901a\u7528\u6027\u3002<\/li>\n
  • \u591a\u573a\u666f\u9002\u914d\u80fd\u529b\uff1a\u540c\u65f6\u652f\u6301\u79bb\u7ebf\u6570\u636e\u7b5b\u9009\uff08\u4e3a\u5c0f\u6a21\u578b\u5fae\u8c03\u7b5b\u9009\u9ad8\u8d28\u91cf\u8f68\u8ff9-\u54cd\u5e94\u6570\u636e\uff09\u3001\u5728\u7ebf\u5f3a\u5316\u5b66\u4e60\uff08\u4e3aGRPO\u7b49\u7b97\u6cd5\u63d0\u4f9b\u7ec6\u7c92\u5ea6\u5956\u52b1\uff09\u548c\u6d4b\u8bd5\u65f6\u7f29\u653e\uff08\u901a\u8fc7\u5956\u52b1\u5f15\u5bfc\u9009\u62e9\u6700\u4f18\u63a8\u7406\u7ed3\u679c\uff09\u3002<\/li>\n
  • \u6570\u636e\u6548\u7387\u4e0e\u6027\u80fd\u5e73\u8861\uff1a7B\u89c4\u6a21\u6a21\u578b\u8d85\u8d8a\u66f4\u5927\u53c2\u6570\u7684\u73b0\u6709PRM\u548c\u4eba\u5de5\u7cbe\u9009\u6570\u636e\uff0c1.5B\u7248\u672c\u9002\u914d\u8fb9\u7f18\u90e8\u7f72\uff0c\u5728\u964d\u4f4e\u8ba1\u7b97\u6210\u672c\u7684\u540c\u65f6\u4fdd\u6301\u4f18\u5f02\u6027\u80fd\u3002<\/li>\n
    \n

    <\/h2>\n","protected":false},"excerpt":{"rendered":"

    \u6587\u7ae0\u6838\u5fc3\u603b\u7ed3\u4e0e\u521b\u65b0\u70b9
    \n\u4e3b\u8981\u5185\u5bb9
    \n\u672c\u6587\u9488\u5bf9\u73b0\u6709\u8fc7\u7a0b\u5956\u52b1\u6a21\u578b\uff08PRMs\uff09\u96be\u4ee5\u8bc4\u4f30\u5927\u8bed\u8a00\u6a21\u578b\u4e2d\u95f4\u63a8\u7406\u8f68\u8ff9\u7684\u95ee\u9898\uff0c\u63d0\u51fa\u4e86\u8f68\u8ff9\u611f\u77e5\u7684PRM\u2014\u2014ReasonFlux-PRM\u3002\u8be5\u6a21\u578b\u901a\u8fc7\u6b65\u7ea7\u548c\u8f68\u8ff9\u7ea7\u53cc\u91cd\u76d1\u7763\uff0c\u9002\u914d\u201c\u8f68\u8ff9-\u54cd\u5e94\u201d\u683c\u5f0f\u7684\u957f\u94fe\u63a8\u7406\u6570\u636e\uff0c\u53ef\u5e94\u7528\u4e8e\u79bb\u7ebf\u9ad8\u8d28\u91cf\u6570\u636e\u7b5b\u9009\u3001\u5728\u7ebf\u5f3a\u5316\u5b66\u4e60\u7b56\u7565\u4f18\u5316\u548c\u6d4b\u8bd5\u65f6Best-of-N\u7f29\u653e\u4e09\u5927\u573a\u666f\u3002\u5728AIME\u3001MATH500\u7b49\u591a\u4e2a\u6311\u6218\u6027\u57fa\u51c6\u6d4b\u8bd5\u4e2d\uff0cReasonFlux-PRM\uff087B\u7248\u672c\uff09\u8868\u73b0\u4f18\u4e8eQwen2.5-Math-PRM-72B\u7b49\u5f3a\u57fa\u7ebf\u6a21\u578b\u548c\u4eba\u5de5<\/p>\n","protected":false},"author":2,"featured_media":76860,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[50,51],"topic":[],"class_list":["post-76861","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-server","tag-50","tag-51"],"yoast_head":"\n2025_NIPS_ReasonFlux-PRM: Trajectory-Aware PRMs for Long Chain-of-Thought Reasoning in LLMs - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/www.wsisp.com\/helps\/76861.html\" \/>\n<meta property=\"og:locale\" content=\"zh_CN\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"2025_NIPS_ReasonFlux-PRM: Trajectory-Aware PRMs for Long Chain-of-Thought Reasoning in LLMs - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\" \/>\n<meta property=\"og:description\" content=\"\u6587\u7ae0\u6838\u5fc3\u603b\u7ed3\u4e0e\u521b\u65b0\u70b9 \u4e3b\u8981\u5185\u5bb9 \u672c\u6587\u9488\u5bf9\u73b0\u6709\u8fc7\u7a0b\u5956\u52b1\u6a21\u578b\uff08PRMs\uff09\u96be\u4ee5\u8bc4\u4f30\u5927\u8bed\u8a00\u6a21\u578b\u4e2d\u95f4\u63a8\u7406\u8f68\u8ff9\u7684\u95ee\u9898\uff0c\u63d0\u51fa\u4e86\u8f68\u8ff9\u611f\u77e5\u7684PRM\u2014\u2014ReasonFlux-PRM\u3002\u8be5\u6a21\u578b\u901a\u8fc7\u6b65\u7ea7\u548c\u8f68\u8ff9\u7ea7\u53cc\u91cd\u76d1\u7763\uff0c\u9002\u914d\u201c\u8f68\u8ff9-\u54cd\u5e94\u201d\u683c\u5f0f\u7684\u957f\u94fe\u63a8\u7406\u6570\u636e\uff0c\u53ef\u5e94\u7528\u4e8e\u79bb\u7ebf\u9ad8\u8d28\u91cf\u6570\u636e\u7b5b\u9009\u3001\u5728\u7ebf\u5f3a\u5316\u5b66\u4e60\u7b56\u7565\u4f18\u5316\u548c\u6d4b\u8bd5\u65f6Best-of-N\u7f29\u653e\u4e09\u5927\u573a\u666f\u3002\u5728AIME\u3001MATH500\u7b49\u591a\u4e2a\u6311\u6218\u6027\u57fa\u51c6\u6d4b\u8bd5\u4e2d\uff0cReasonFlux-PRM\uff087B\u7248\u672c\uff09\u8868\u73b0\u4f18\u4e8eQwen2.5-Math-PRM-72B\u7b49\u5f3a\u57fa\u7ebf\u6a21\u578b\u548c\u4eba\u5de5\" \/>\n<meta property=\"og:url\" content=\"https:\/\/www.wsisp.com\/helps\/76861.html\" \/>\n<meta property=\"og:site_name\" content=\"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\" \/>\n<meta property=\"article:published_time\" content=\"2026-02-23T06:07:37+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/www.wsisp.com\/helps\/wp-content\/uploads\/2026\/02\/20260223060736-699beea8352cf.png\" \/>\n<meta name=\"author\" content=\"admin\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"\u4f5c\u8005\" \/>\n\t<meta name=\"twitter:data1\" content=\"admin\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/76861.html\",\"url\":\"https:\/\/www.wsisp.com\/helps\/76861.html\",\"name\":\"2025_NIPS_ReasonFlux-PRM: Trajectory-Aware PRMs for Long Chain-of-Thought Reasoning in LLMs - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\",\"isPartOf\":{\"@id\":\"https:\/\/www.wsisp.com\/helps\/#website\"},\"datePublished\":\"2026-02-23T06:07:37+00:00\",\"dateModified\":\"2026-02-23T06:07:37+00:00\",\"author\":{\"@id\":\"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41\"},\"breadcrumb\":{\"@id\":\"https:\/\/www.wsisp.com\/helps\/76861.html#breadcrumb\"},\"inLanguage\":\"zh-Hans\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/www.wsisp.com\/helps\/76861.html\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/76861.html#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"\u9996\u9875\",\"item\":\"https:\/\/www.wsisp.com\/helps\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"2025_NIPS_ReasonFlux-PRM: Trajectory-Aware PRMs for Long Chain-of-Thought Reasoning in LLMs\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/#website\",\"url\":\"https:\/\/www.wsisp.com\/helps\/\",\"name\":\"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\",\"description\":\"\u9999\u6e2f\u670d\u52a1\u5668_\u9999\u6e2f\u4e91\u670d\u52a1\u5668\u8d44\u8baf_\u670d\u52a1\u5668\u5e2e\u52a9\u6587\u6863_\u670d\u52a1\u5668\u6559\u7a0b\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/www.wsisp.com\/helps\/?s={search_term_string}\"},\"query-input\":\"required name=search_term_string\"}],\"inLanguage\":\"zh-Hans\"},{\"@type\":\"Person\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41\",\"name\":\"admin\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"zh-Hans\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery\",\"contentUrl\":\"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery\",\"caption\":\"admin\"},\"sameAs\":[\"http:\/\/wp.wsisp.com\"],\"url\":\"https:\/\/www.wsisp.com\/helps\/author\/admin\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"2025_NIPS_ReasonFlux-PRM: Trajectory-Aware PRMs for Long Chain-of-Thought Reasoning in LLMs - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/www.wsisp.com\/helps\/76861.html","og_locale":"zh_CN","og_type":"article","og_title":"2025_NIPS_ReasonFlux-PRM: Trajectory-Aware PRMs for Long Chain-of-Thought Reasoning in LLMs - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","og_description":"\u6587\u7ae0\u6838\u5fc3\u603b\u7ed3\u4e0e\u521b\u65b0\u70b9 \u4e3b\u8981\u5185\u5bb9 \u672c\u6587\u9488\u5bf9\u73b0\u6709\u8fc7\u7a0b\u5956\u52b1\u6a21\u578b\uff08PRMs\uff09\u96be\u4ee5\u8bc4\u4f30\u5927\u8bed\u8a00\u6a21\u578b\u4e2d\u95f4\u63a8\u7406\u8f68\u8ff9\u7684\u95ee\u9898\uff0c\u63d0\u51fa\u4e86\u8f68\u8ff9\u611f\u77e5\u7684PRM\u2014\u2014ReasonFlux-PRM\u3002\u8be5\u6a21\u578b\u901a\u8fc7\u6b65\u7ea7\u548c\u8f68\u8ff9\u7ea7\u53cc\u91cd\u76d1\u7763\uff0c\u9002\u914d\u201c\u8f68\u8ff9-\u54cd\u5e94\u201d\u683c\u5f0f\u7684\u957f\u94fe\u63a8\u7406\u6570\u636e\uff0c\u53ef\u5e94\u7528\u4e8e\u79bb\u7ebf\u9ad8\u8d28\u91cf\u6570\u636e\u7b5b\u9009\u3001\u5728\u7ebf\u5f3a\u5316\u5b66\u4e60\u7b56\u7565\u4f18\u5316\u548c\u6d4b\u8bd5\u65f6Best-of-N\u7f29\u653e\u4e09\u5927\u573a\u666f\u3002\u5728AIME\u3001MATH500\u7b49\u591a\u4e2a\u6311\u6218\u6027\u57fa\u51c6\u6d4b\u8bd5\u4e2d\uff0cReasonFlux-PRM\uff087B\u7248\u672c\uff09\u8868\u73b0\u4f18\u4e8eQwen2.5-Math-PRM-72B\u7b49\u5f3a\u57fa\u7ebf\u6a21\u578b\u548c\u4eba\u5de5","og_url":"https:\/\/www.wsisp.com\/helps\/76861.html","og_site_name":"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","article_published_time":"2026-02-23T06:07:37+00:00","og_image":[{"url":"https:\/\/www.wsisp.com\/helps\/wp-content\/uploads\/2026\/02\/20260223060736-699beea8352cf.png"}],"author":"admin","twitter_card":"summary_large_image","twitter_misc":{"\u4f5c\u8005":"admin"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/www.wsisp.com\/helps\/76861.html","url":"https:\/\/www.wsisp.com\/helps\/76861.html","name":"2025_NIPS_ReasonFlux-PRM: Trajectory-Aware PRMs for Long Chain-of-Thought Reasoning in LLMs - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","isPartOf":{"@id":"https:\/\/www.wsisp.com\/helps\/#website"},"datePublished":"2026-02-23T06:07:37+00:00","dateModified":"2026-02-23T06:07:37+00:00","author":{"@id":"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41"},"breadcrumb":{"@id":"https:\/\/www.wsisp.com\/helps\/76861.html#breadcrumb"},"inLanguage":"zh-Hans","potentialAction":[{"@type":"ReadAction","target":["https:\/\/www.wsisp.com\/helps\/76861.html"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/www.wsisp.com\/helps\/76861.html#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"\u9996\u9875","item":"https:\/\/www.wsisp.com\/helps"},{"@type":"ListItem","position":2,"name":"2025_NIPS_ReasonFlux-PRM: Trajectory-Aware PRMs for Long Chain-of-Thought Reasoning in LLMs"}]},{"@type":"WebSite","@id":"https:\/\/www.wsisp.com\/helps\/#website","url":"https:\/\/www.wsisp.com\/helps\/","name":"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","description":"\u9999\u6e2f\u670d\u52a1\u5668_\u9999\u6e2f\u4e91\u670d\u52a1\u5668\u8d44\u8baf_\u670d\u52a1\u5668\u5e2e\u52a9\u6587\u6863_\u670d\u52a1\u5668\u6559\u7a0b","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/www.wsisp.com\/helps\/?s={search_term_string}"},"query-input":"required name=search_term_string"}],"inLanguage":"zh-Hans"},{"@type":"Person","@id":"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41","name":"admin","image":{"@type":"ImageObject","inLanguage":"zh-Hans","@id":"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/image\/","url":"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery","contentUrl":"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery","caption":"admin"},"sameAs":["http:\/\/wp.wsisp.com"],"url":"https:\/\/www.wsisp.com\/helps\/author\/admin"}]}},"_links":{"self":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/posts\/76861","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/comments?post=76861"}],"version-history":[{"count":0,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/posts\/76861\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/media\/76860"}],"wp:attachment":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/media?parent=76861"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/categories?post=76861"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/tags?post=76861"},{"taxonomy":"topic","embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/topic?post=76861"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}