{"id":14087,"date":"2026-10-27T16:58:12","date_gmt":"2026-10-27T08:58:12","guid":{"rendered":"https:\/\/infernews.com\/blog\/robojudge\/"},"modified":"2026-10-27T16:58:12","modified_gmt":"2026-10-27T08:58:12","slug":"robojudge","status":"publish","type":"post","link":"https:\/\/infernews.com\/blog\/robojudge\/","title":{"rendered":"RoboJudge\uff1a\u628a\u5f71\u7247\u7684\u7269\u7406\u5408\u7406\u6027\u8207\u6307\u4ee4\u5206\u958b\u8a55\u6e2c"},"content":{"rendered":"\n<figure class=\"wp-block-image size-large\"><img decoding=\"async\" src=\"https:\/\/infernews.com\/blog\/wp-content\/uploads\/2026\/10\/RoboJudge.jpg\" alt=\"RoboJudge overview\"><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">\u5177\u8eab\u5f71\u7247\u53ef\u80fd\u52d5\u4f5c\u81ea\u7136\u537b\u672a\u5b8c\u6210\u6307\u4ee4\uff0c\u4e5f\u53ef\u80fd\u5b8c\u6210\u4efb\u52d9\u4f46\u9055\u53cd\u7269\u7406\u898f\u5f8b\u3002RoboJudge \u4ee5 Multimodal Language Models (MLLMs) \u8a55\u5be9\u751f\u6210\u5f71\u7247\uff0c\u5206\u958b\u8861\u91cf Physical Adherence (PA\uff0c\u7269\u7406\u9075\u5faa) \u8207 Instruction Alignment (IA\uff0c\u6307\u4ee4\u7b26\u5408\u5ea6)\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u9019\u6709\u52a9\u6a5f\u68b0\u4eba\u53ca\u5177\u8eab\u4e16\u754c\u6a21\u578b\u5718\u968a\u5206\u8fa8\u52d5\u4f5c\u53ef\u4fe1\u5ea6\u548c\u4efb\u52d9\u5b8c\u6210\u5ea6\u3002RoboJudge-Bench \u6536\u9304 800 \u6bb5\u4eba\u5de5\u6a19\u8a3b\u5f71\u7247\uff0c\u6db5\u84cb 8 \u500b\u8a9e\u6599\u4f86\u6e90\u300110 \u500b\u5f71\u7247\u751f\u6210\u5668\uff0c\u4ee5\u53ca\u593e\u722a\u3001\u4eba\u624b\u8207\u9748\u5de7\u624b\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">RoboJudge-Train \u6536\u9304 12,000 \u6bb5\u5f71\u7247\u53ca 25,000 \u7b46\u4eba\u5de5\u8a55\u5206\u8207\u7406\u7531\uff1bRoboJudge-9B \u7d93\u76e3\u7763\u5b78\u7fd2\u53ca\u5f37\u5316\u5b78\u7fd2\u8a13\u7df4\u3002Pearson \u76f8\u95dc\u4fc2\u6578\u5728 PA \u70ba 0.645\u3001IA \u70ba 0.775\u3001\u6574\u9ad4\u70ba 0.719\uff0c\u8f03\u6e2c\u8a66\u4e2d\u7684\u6700\u4f73\u5c08\u6709\u57fa\u7dda\u9ad8 0.089\uff1b\u76f8\u95dc\u6027\u4e0d\u4ee3\u8868\u6bcf\u6b21\u5224\u65b7\u90fd\u6b63\u78ba\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">GitHub \u63d0\u4f9b\u63a8\u7406\u548c\u57fa\u6e96\u8a55\u4f30\u7a0b\u5f0f\uff0c\u4e26\u9023\u7d50\u6a21\u578b\u8207\u8cc7\u6599\u96c6\u3002README \u5217\u51fa\u4f9d\u8cf4\u5b89\u88dd\u53ca\u57f7\u884c\u6d41\u7a0b\uff0c\u4f46\u6c92\u6709\u786c\u4ef6\u9700\u6c42\u8cc7\u6599\uff0c\u63a1\u7528\u524d\u5b9c\u5148\u78ba\u8a8d\u8a2d\u5099\u80fd\u5426\u8ca0\u64d4\u6a21\u578b\u8f09\u5165\u8207\u5f71\u7247\u8655\u7406\u3002<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>PA \u6aa2\u8996\u5f71\u7247\u662f\u5426\u7b26\u5408\u7269\u7406\u8207\u6642\u9593\u7d04\u675f\u3002<\/li>\n<li>IA \u6aa2\u8996\u5f71\u7247\u52d5\u4f5c\u662f\u5426\u7b26\u5408\u6307\u4ee4\u3002<\/li>\n<li>\u57fa\u6e96\u53ef\u7528\u65bc\u6bd4\u8f03\u8a55\u5206\uff0c\u4e0d\u80fd\u53d6\u4ee3\u4eba\u5de5\u6838\u5c0d\u3002<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/spherelab.ai\/RoboJudge\/\" rel=\"noopener noreferrer\"><strong>\u9805\u76ee\u4e3b\u9801<\/strong><\/a> \u00b7 <a href=\"https:\/\/github.com\/Sphere-AI-Lab\/RoboJudge\" rel=\"noopener noreferrer\"><strong>GitHub<\/strong><\/a> \u00b7 <a href=\"https:\/\/huggingface.co\/HuggingFriends\/RoboJudge-9B\" rel=\"noopener noreferrer\"><strong>\u6a21\u578b<\/strong><\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>RoboJudge \u5206\u958b\u6aa2\u8996\u5177\u8eab\u5f71\u7247\u7684\u7269\u7406\u5408\u7406\u6027\u8207\u6307\u4ee4\u5b8c\u6210\u5ea6\uff0c\u4e26\u4ee5\u4eba\u985e\u8a55\u5206\u8a13\u7df4\u6a21\u578b\uff0c\u4f9b\u5718\u968a\u6bd4\u8f03\u751f\u6210\u7d50\u679c\u3002<\/p>\n","protected":false},"author":8,"featured_media":14086,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"wpai_generated_summary":"","wpai_meta_description":"","footnotes":""},"categories":[133,164,162,76,119,186,127,31,163,199,214],"tags":[],"class_list":["post-14087","post","type-post","status-publish","format-standard","hentry","category-133","category-164","category-ai-productions","category-76","category-119","category-186","category-127","category-video","category-163","category-dataset-","category-reinforcement-learning"],"_links":{"self":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts\/14087","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=14087"}],"version-history":[{"count":0,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts\/14087\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/media\/14086"}],"wp:attachment":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/media?parent=14087"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/categories?post=14087"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/tags?post=14087"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}