{"id":35227,"date":"2024-12-21T03:59:50","date_gmt":"2024-12-20T19:59:50","guid":{"rendered":"https:\/\/17aitech.com\/?p=35227"},"modified":"2024-12-21T03:59:50","modified_gmt":"2024-12-20T19:59:50","slug":"%e7%94%a8ai%e8%87%aa%e5%8a%a8%e8%ae%be%e8%ae%a1%e6%99%ba%e8%83%bd%e4%bd%93%ef%bc%8c%e6%95%b0%e5%ad%a6%e6%8f%90%e5%88%8625-9%ef%bc%8c%e8%bf%9c%e8%b6%85%e6%89%8b%e5%b7%a5%e8%ae%be%e8%ae%a1","status":"publish","type":"post","link":"https:\/\/17aitech.com\/?p=35227","title":{"rendered":"\u7528AI\u81ea\u52a8\u8bbe\u8ba1\u667a\u80fd\u4f53\uff0c\u6570\u5b66\u63d0\u520625.9%\uff0c\u8fdc\u8d85\u624b\u5de5\u8bbe\u8ba1"},"content":{"rendered":"<p>\u6587\u7ae0\u6765\u6e90\u4e8e\u4e92\u8054\u7f51:<a href=\"https:\/\/www.jiqizhixin.com\/articles\/2024-08-22-5\" target=\"_blank\">\u7528AI\u81ea\u52a8\u8bbe\u8ba1\u667a\u80fd\u4f53\uff0c\u6570\u5b66\u63d0\u520625.9%\uff0c\u8fdc\u8d85\u624b\u5de5\u8bbe\u8ba1<\/a><\/p>\n<blockquote data-author-name=\"\" data-content-utf8-length=\"33\" data-source-title=\"\" data-type=\"2\" data-url=\"\">\n<section>\n<section>\n<section>\n<p>\u57fa\u4e8e ADAS \u6240\u53d1\u73b0\u7684\u667a\u80fd\u4f53\u7684\u6027\u80fd\u5927\u5927\u4f18\u4e8e\u6700\u5148\u8fdb\u7684\u624b\u5de5\u8bbe\u8ba1\u7684\u57fa\u7ebf\u3002<\/p>\n<\/section>\n<\/section>\n<\/section>\n<\/blockquote>\n<p>\u57fa\u7840\u6a21\u578b (FM) \u5982 GPT \u548c Claude \uff0c\u6b63\u5728\u6210\u4e3a\u901a\u7528\u667a\u80fd\u4f53\u7684\u5f3a\u6709\u529b\u652f\u6301\uff0c\u88ab\u8d8a\u6765\u8d8a\u591a\u7684\u7528\u4e8e\u591a\u79cd\u63a8\u7406\u548c\u89c4\u5212\u4efb\u52a1\u3002<\/p>\n<p>\u7136\u800c\uff0c\u5728\u89e3\u51b3\u95ee\u9898\u65f6\uff0c\u9700\u8981\u7684\u667a\u80fd\u4f53\u901a\u5e38\u662f\u5177\u6709\u591a\u4e2a\u7ec4\u4ef6\u7684\u590d\u5408\u667a\u80fd\u4f53\u7cfb\u7edf\uff0c\u800c\u4e0d\u662f\u5355\u7247\u6a21\u578b\u67e5\u8be2 \u3002\u6b64\u5916\uff0c\u4e3a\u4e86\u4f7f\u667a\u80fd\u4f53\u80fd\u591f\u89e3\u51b3\u590d\u6742\u7684\u73b0\u5b9e\u4e16\u754c\u4efb\u52a1\uff0c\u5b83\u4eec\u901a\u5e38\u9700\u8981\u8bbf\u95ee\u5916\u90e8\u5de5\u5177\uff0c\u4f8b\u5982\u641c\u7d22\u5f15\u64ce\u3001\u4ee3\u7801\u6267\u884c\u548c\u6570\u636e\u5e93\u67e5\u8be2\u3002<\/p>\n<p>\u56e0\u6b64\uff0c\u4eba\u4eec\u63d0\u51fa\u4e86\u8bb8\u591a\u6709\u6548\u7684\u667a\u80fd\u4f53\u7cfb\u7edf\u6784\u5efa\u5757\uff0c\u4f8b\u5982\u601d\u7ef4\u94fe\u89c4\u5212\u548c\u63a8\u7406\u3001\u8bb0\u5fc6\u7ed3\u6784\u3001\u5de5\u5177\u4f7f\u7528\u548c\u81ea\u6211\u53cd\u601d\u3002\u5c3d\u7ba1\u8fd9\u4e9b\u667a\u80fd\u4f53\u5df2\u7ecf\u5728\u5404\u79cd\u5e94\u7528\u4e2d\u53d6\u5f97\u4e86\u663e\u8457\u7684\u6210\u529f\uff0c\u4f46\u5f00\u53d1\u8fd9\u4e9b\u6784\u5efa\u5757\u5e76\u5c06\u5b83\u4eec\u7ec4\u5408\u6210\u590d\u6742\u7684\u667a\u80fd\u4f53\u7cfb\u7edf\u901a\u5e38\u9700\u8981\u7279\u5b9a\u9886\u57df\u7684\u624b\u52a8\u8c03\u6574\u4ee5\u53ca\u7814\u7a76\u4eba\u5458\u548c\u5de5\u7a0b\u5e08\u7684\u5927\u91cf\u52aa\u529b\u3002<\/p>\n<p>\u7136\u800c\uff0c\u673a\u5668\u5b66\u4e60\u7684\u5386\u53f2\u544a\u8bc9\u6211\u4eec\uff0c\u624b\u5de5\u8bbe\u8ba1\u7684\u89e3\u51b3\u65b9\u6848\u6700\u7ec8\u4f1a\u88ab\u6a21\u578b\u5b66\u4e60\u5230\u7684\u89e3\u51b3\u65b9\u6848\u6240\u53d6\u4ee3\u3002<\/p>\n<p>\u672c\u6587\uff0c\u6765\u81ea\u4e0d\u5217\u98a0\u54e5\u4f26\u6bd4\u4e9a\u5927\u5b66\u3001\u975e\u8425\u5229\u4eba\u5de5\u667a\u80fd\u7814\u7a76\u673a\u6784 Vector Institute \u7b49\u7684\u7814\u7a76\u8005\u5236\u5b9a\u4e86\u4e00\u4e2a\u65b0\u7684\u7814\u7a76\u9886\u57df\uff0c\u5373\u667a\u80fd\u4f53\u7cfb\u7edf\u7684\u81ea\u52a8\u5316\u8bbe\u8ba1\uff08ADAS\uff0cAutomated Design of Agentic Systems\uff09\uff0c\u5e76\u63d0\u51fa\u4e86\u4e00\u79cd\u7b80\u5355\u4f46\u6709\u6548\u7684 ADAS \u7b97\u6cd5\uff0c\u540d\u4e3a\u5143\u667a\u80fd\u4f53\u641c\u7d22\uff08Meta Agent Search\uff09\uff0c\u4ee5\u8bc1\u660e\u667a\u80fd\u4f53\u53ef\u4ee5\u901a\u8fc7\u4ee3\u7801\u7f16\u7a0b\u6765\u53d1\u660e\u65b0\u9896\u800c\u5f3a\u5927\u7684\u667a\u80fd\u4f53\u8bbe\u8ba1\u3002<\/p>\n<p>\u8be5\u7814\u7a76\u65e8\u5728\u81ea\u52a8\u521b\u5efa\u5f3a\u5927\u7684\u667a\u80fd\u4f53\u7cfb\u7edf\u8bbe\u8ba1\uff0c\u5305\u62ec\u5f00\u53d1\u65b0\u7684\u6784\u5efa\u5757\u5e76\u4ee5\u65b0\u7684\u65b9\u5f0f\u7ec4\u5408\u5b83\u4eec\u3002<\/p>\n<p>\u5b9e\u9a8c\u8868\u660e\uff0c\u57fa\u4e8e ADAS \u6240\u53d1\u73b0\u7684\u667a\u80fd\u4f53\u7684\u6027\u80fd\u5927\u5927\u4f18\u4e8e\u6700\u5148\u8fdb\u7684\u624b\u5de5\u8bbe\u8ba1\u7684\u57fa\u7ebf\u3002\u4f8b\u5982\uff0c\u672c\u6587\u8bbe\u8ba1\u7684\u667a\u80fd\u4f53\u5728 DROP \u7684\u9605\u8bfb\u7406\u89e3\u4efb\u52a1\u4e2d\u5c06 F1 \u5206\u6570\u63d0\u9ad8\u4e86 13.6\/100\uff08\u4e0e\u57fa\u7ebf\u6bd4\uff09\uff0c\u5728 MGSM \u7684\u6570\u5b66\u4efb\u52a1\u4e2d\u5c06\u51c6\u786e\u7387\u63d0\u9ad8\u4e86 14.4%\u3002\u6b64\u5916\uff0c\u5728\u8de8\u57df\u8fc1\u79fb\u540e\uff0c\u5b83\u4eec\u5728 GSM8K \u548c GSM-Hard \u6570\u5b66\u4efb\u52a1\u4e0a\u7684\u51c6\u786e\u7387\u5206\u522b\u6bd4\u57fa\u7ebf\u63d0\u9ad8\u4e86 25.9% \u548c 13.2%\u3002<\/p>\n<p>\u4e0e\u624b\u5de5\u8bbe\u8ba1\u7684\u89e3\u51b3\u65b9\u6848\u76f8\u6bd4\uff0c\u672c\u6587\u7b97\u6cd5\u8868\u73b0\u51fa\u8272\uff0c\u8fd9\u8bf4\u660e ADAS \u5728\u81ea\u52a8\u5316\u667a\u80fd\u4f53\u7cfb\u7edf\u8bbe\u8ba1\u65b9\u9762\u7684\u6f5c\u529b\u3002\u6b64\u5916\uff0c\u5b9e\u9a8c\u8868\u660e\uff0c\u6240\u53d1\u73b0\u7684\u667a\u80fd\u4f53\u4e0d\u4ec5\u5728\u8de8\u76f8\u4f3c\u9886\u57df\u8fc1\u79fb\u65f6\u8868\u73b0\u826f\u597d\uff0c\u800c\u4e14\u5728\u8de8\u4e0d\u540c\u9886\u57df\u8fc1\u79fb\u65f6\u4e5f\u8868\u73b0\u51fa\u8272\uff0c\u4f8b\u5982\u4ece\u6570\u5b66\u5230\u9605\u8bfb\u7406\u89e3\u3002<\/p>\n<p><a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/08\/frc-072da4cb6fbf156afe1e1191a8b7d99d.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/08\/frc-072da4cb6fbf156afe1e1191a8b7d99d.png\"><\/a><\/p>\n<ul>\n<li>\n<p>\u8bba\u6587\u5730\u5740\uff1ahttps:\/\/arxiv.org\/pdf\/2408.08435<\/p>\n<\/li>\n<li>\n<p>\u9879\u76ee\u5730\u5740\uff1ahttps:\/\/github.com\/ShengranHu\/ADAS<\/p>\n<\/li>\n<li>\n<p>\u8bba\u6587\u4e3b\u9875\uff1ahttps:\/\/www.shengranhu.com\/ADAS\/<\/p>\n<\/li>\n<li>\n<p>\u8bba\u6587\u6807\u9898\uff1aAutomated Design of Agentic Systems<\/p>\n<\/li>\n<\/ul>\n<p><strong>\u65b0\u7814\u7a76\u9886\u57df\uff1a\u667a\u80fd\u4f53\u7cfb\u7edf\u7684\u81ea\u52a8\u5316\u8bbe\u8ba1\uff08ADAS\uff09<\/strong><\/p>\n<p>\u8be5\u7814\u7a76\u63d0\u51fa\u4e86\u4e00\u4e2a\u65b0\u7684\u7814\u7a76\u9886\u57df \u2014\u2014 \u667a\u80fd\u4f53\u7cfb\u7edf\u7684\u81ea\u52a8\u5316\u8bbe\u8ba1\uff08Automated Design of Agentic Systems\uff0cADAS\uff09\uff0c\u5e76\u63cf\u8ff0\u4e86 ADAS \u7b97\u6cd5\u7684\u4e09\u4e2a\u5173\u952e\u7ec4\u6210\u90e8\u5206\u2014\u2014\u641c\u7d22\u7a7a\u95f4\u3001\u641c\u7d22\u7b97\u6cd5\u3001\u8bc4\u4f30\u51fd\u6570\u3002ADAS \u4f7f\u7528\u641c\u7d22\u7b97\u6cd5\u8de8\u641c\u7d22\u7a7a\u95f4\u6765\u53d1\u73b0\u667a\u80fd\u4f53\u7cfb\u7edf\u3002<\/p>\n<p><a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/08\/frc-81eab76f9a18c00fb13356139ce33766.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/08\/frc-81eab76f9a18c00fb13356139ce33766.png\"><\/a><\/p>\n<p><strong>\u641c\u7d22\u7a7a\u95f4<\/strong>\uff1a\u641c\u7d22\u7a7a\u95f4\u5b9a\u4e49\u4e86\u54ea\u4e9b\u667a\u80fd\u4f53\u7cfb\u7edf\u53ef\u4ee5\u5728 ADAS \u4e2d\u88ab\u8868\u5f81\u5e76\u88ab\u53d1\u73b0\u3002\u4f8b\u5982\uff0c\u50cf PromptBreeder (Fernando et al., 2024) \u8fd9\u6837\u7684\u5de5\u4f5c\u4ec5\u6539\u53d8\u667a\u80fd\u4f53\u7684\u6587\u672c\u63d0\u793a\uff0c\u800c\u5176\u4ed6\u7ec4\u4ef6\uff08\u4f8b\u5982\u63a7\u5236\u6d41\uff09\u4fdd\u6301\u4e0d\u53d8\u3002\u56e0\u6b64\uff0c\u5728\u641c\u7d22\u7a7a\u95f4\u4e2d\uff0c\u65e0\u6cd5\u8868\u5f81\u5177\u6709\u4e0e\u9884\u5b9a\u4e49\u63a7\u5236\u6d41\u4e0d\u540c\u63a7\u5236\u6d41\u7684\u667a\u80fd\u4f53\u3002<\/p>\n<p><strong>\u641c\u7d22\u7b97\u6cd5<\/strong>\uff1a\u641c\u7d22\u7b97\u6cd5\u5b9a\u4e49\u4e86 ADAS \u7b97\u6cd5\u5982\u4f55\u63a2\u7d22\u641c\u7d22\u7a7a\u95f4\u3002\u7531\u4e8e\u641c\u7d22\u7a7a\u95f4\u901a\u5e38\u975e\u5e38\u5927\u751a\u81f3\u65e0\u754c\uff0c\u56e0\u6b64\u5e94\u8003\u8651\u63a2\u7d22\u4e0e\u5229\u7528\u7684\u6743\u8861\uff08Sutton &amp; Barto\uff0c2018\uff09\u3002\u7406\u60f3\u60c5\u51b5\u4e0b\uff0c\u8be5\u7b97\u6cd5\u65e2\u53ef\u4ee5\u5feb\u901f\u53d1\u73b0\u9ad8\u6027\u80fd\u667a\u80fd\u4f53\u7cfb\u7edf\uff0c\u53c8\u53ef\u4ee5\u907f\u514d\u9677\u5165\u5c40\u90e8\u6700\u4f18\u3002\u73b0\u6709\u65b9\u6cd5\u5305\u62ec\u4f7f\u7528\u5f3a\u5316\u5b66\u4e60\uff08Zhuge et al., 2024\uff09\u6216\u8fed\u4ee3\u751f\u6210\u65b0\u89e3\u51b3\u65b9\u6848\u7684 FM\uff08Fernando et al., 2024\uff09\u4f5c\u4e3a\u641c\u7d22\u7b97\u6cd5\u3002<\/p>\n<p><strong>\u8bc4\u4f30\u51fd\u6570<\/strong>\uff1a\u6839\u636e ADAS \u7b97\u6cd5\u7684\u5e94\u7528\uff0c\u53ef\u80fd\u9700\u8981\u8003\u8651\u4e0d\u540c\u7684\u4f18\u5316\u76ee\u6807\uff0c\u4f8b\u5982\u667a\u80fd\u4f53\u7684\u6027\u80fd\u3001\u6210\u672c\u3001\u5ef6\u8fdf\u6216\u5b89\u5168\u6027\u3002\u8bc4\u4f30\u51fd\u6570\u5b9a\u4e49\u5982\u4f55\u8bc4\u4f30\u5019\u9009\u667a\u80fd\u4f53\u7684\u8fd9\u4e9b\u6307\u6807\u3002\u4f8b\u5982\uff0c\u4e3a\u4e86\u8bc4\u4f30\u667a\u80fd\u4f53\u5728\u672a\u89c1\u8fc7\u7684\u6570\u636e\u4e0a\u7684\u8868\u73b0\uff0c\u4e00\u79cd\u7b80\u5355\u7684\u65b9\u6cd5\u662f\u8ba1\u7b97\u4efb\u52a1\u9a8c\u8bc1\u6570\u636e\u7684\u51c6\u786e\u7387\u3002<\/p>\n<p>\u8be5\u7814\u7a76\u63d0\u51fa\u7684\u7b80\u5355\u4f46\u6709\u6548\u7684 ADAS \u7b97\u6cd5\u2014\u2014\u5143\u667a\u80fd\u4f53\u641c\u7d22\u7684\u6838\u5fc3\u6982\u5ff5\u662f\u6307\u793a\u5143\u667a\u80fd\u4f53\u8fed\u4ee3\u5730\u521b\u5efa\u6709\u8da3\u7684\u65b0\u667a\u80fd\u4f53\uff0c\u8bc4\u4f30\u5b83\u4eec\uff0c\u5c06\u5b83\u4eec\u6dfb\u52a0\u5230\u667a\u80fd\u4f53\u5b58\u50a8\u5e93\u4e2d\uff0c\u5e76\u4f7f\u7528\u6b64\u5b58\u50a8\u5e93\u5e2e\u52a9\u5143\u667a\u80fd\u4f53\u5728\u540e\u7eed\u8fed\u4ee3\u4e2d\u521b\u5efa\u66f4\u6709\u8da3\u7684\u65b0\u667a\u80fd\u4f53\u3002\u4e0e\u73b0\u6709\u7684\u5229\u7528\u4eba\u7c7b\u5174\u8da3\u6982\u5ff5\u7684\u5f00\u653e\u5f0f\u7b97\u6cd5\u7c7b\u4f3c\uff0c\u8be5\u7814\u7a76\u9f13\u52b1\u5143\u667a\u80fd\u4f53\u63a2\u7d22\u6709\u8da3\u7684\u3001\u6709\u4ef7\u503c\u7684\u667a\u80fd\u4f53\u3002<\/p>\n<p>\u5143\u667a\u80fd\u4f53\u641c\u7d22\u7684\u6838\u5fc3\u601d\u60f3\u662f\u91c7\u7528 FM \u4f5c\u4e3a\u641c\u7d22\u7b97\u6cd5\uff0c\u57fa\u4e8e\u4e0d\u65ad\u589e\u957f\u7684\u667a\u80fd\u4f53\u5b58\u50a8\u5e93\u6765\u8fed\u4ee3\u7f16\u7a0b\u6709\u8da3\u7684\u65b0\u667a\u80fd\u4f53\u3002\u8be5\u7814\u7a76\u4e3a\u5143\u667a\u80fd\u4f53\u5b9a\u4e49\u4e86\u4e00\u4e2a\u7b80\u5355\u7684\u6846\u67b6\uff08100 \u884c\u4ee3\u7801\u4ee5\u5185\uff09\uff0c\u4e3a\u5176\u63d0\u4f9b\u4e86\u4e00\u7ec4\u57fa\u672c\u529f\u80fd\uff0c\u4f8b\u5982\u67e5\u8be2 FM \u6216\u683c\u5f0f\u5316\u63d0\u793a\u3002<\/p>\n<p>\u56e0\u6b64\uff0c\u5143\u667a\u80fd\u4f53\u53ea\u9700\u8981\u7f16\u5199\u4e00\u4e2a\u300c\u524d\u5411\u300d\u51fd\u6570\u6765\u5b9a\u4e49\u4e00\u4e2a\u65b0\u7684\u667a\u80fd\u4f53\u7cfb\u7edf\uff0c\u7c7b\u4f3c\u4e8e FunSearch \u4e2d\u7684\u505a\u6cd5\uff08Romera-Paredes et al., 2024\uff09\u3002\u8be5\u51fd\u6570\u63a5\u6536\u4efb\u52a1\u4fe1\u606f\u5e76\u8f93\u51fa\u667a\u80fd\u4f53\u5bf9\u4efb\u52a1\u7684\u54cd\u5e94\u3002<\/p>\n<p>\u5982\u56fe 1 \u6240\u793a\uff0c\u5143\u667a\u80fd\u4f53\u641c\u7d22\u7684\u6838\u5fc3\u601d\u60f3\u662f\u8ba9\u5143\u667a\u80fd\u4f53\u5728\u4ee3\u7801\u4e2d\u8fed\u4ee3\u5730\u7f16\u7a0b\u65b0\u7684\u667a\u80fd\u4f53\u3002\u4e0b\u9762\u663e\u793a\u4e86\u5143\u667a\u80fd\u4f53\u7a0b\u5e8f\u65b0\u667a\u80fd\u4f53\u7a0b\u5e8f\u7684\u4e3b\u8981\u63d0\u793a\uff0c\u5176\u4e2d\u63d0\u793a\u4e2d\u7684\u53d8\u91cf\u9ad8\u4eae\u663e\u793a\u3002<\/p>\n<p><a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/08\/frc-057ce16c5d8022318b7ffd0785d2458d.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/08\/frc-057ce16c5d8022318b7ffd0785d2458d.png\"><\/a><\/p>\n<p><a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/08\/frc-1eceaa9ea7e5bbc225f5475c71d65afa.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/08\/frc-1eceaa9ea7e5bbc225f5475c71d65afa.png\"><\/a><\/p>\n<p><strong>\u5b9e\u9a8c<\/strong><\/p>\n<p>\u6240\u6709\u5b9e\u9a8c\u7ed3\u679c\u8868\u660e\u672c\u6587\u53d1\u73b0\u7684\u667a\u80fd\u4f53\u5927\u5927\u4f18\u4e8e\u57fa\u7ebf\u6700\u5148\u8fdb\u7684\u624b\u5de5\u8bbe\u8ba1\u7684\u667a\u80fd\u4f53\u3002\u503c\u5f97\u6ce8\u610f\u7684\u662f\uff0c\u8be5\u7814\u7a76\u53d1\u73b0\u7684\u667a\u80fd\u4f53\u5728 DROP \u9605\u8bfb\u7406\u89e3\u4efb\u52a1\u4e0a\u6bd4\u57fa\u7ebf\u63d0\u9ad8\u4e86 13.6\/100\uff08F1 \u5206\u6570\uff09\uff0c\u5728 MGSM \u6570\u5b66\u4efb\u52a1\u4e0a\u6bd4\u57fa\u7ebf\u63d0\u9ad8\u4e86 14.4%\uff08\u51c6\u786e\u7387\uff09\u3002\u6b64\u5916\uff0c\u7814\u7a76\u8005\u53d1\u73b0\u7684\u667a\u80fd\u4f53\u5728\u4ece GPT-3.5 \u8fc1\u79fb\u5230 GPT-4 \u540e\uff0c\u5728 ARC \u4efb\u52a1\u4e0a\u7684\u8868\u73b0\u6bd4\u57fa\u7ebf\u63d0\u9ad8\u4e86 14%\uff08\u51c6\u786e\u7387\uff09\uff0c\u5728\u4ece MGSM \u6570\u5b66\u4efb\u52a1\u8fc1\u79fb\u5230 GSM8K \u548c GSM-Hard \u4e2d\u7684 held-out \u6570\u5b66\u4efb\u52a1\u540e\uff0c\u51c6\u786e\u7387\u5206\u522b\u63d0\u9ad8\u4e86 25.9% \u548c 13.2%\u3002<\/p>\n<p><strong>\u6848\u4f8b\u7814\u7a76\uff1aARC \u6311\u6218<\/strong><\/p>\n<p>\u5982\u56fe 3a \u6240\u793a\uff0c\u5143\u667a\u80fd\u4f53\u641c\u7d22\u53ef\u4ee5\u6709\u6548\u4e14\u9010\u6b65\u5730\u53d1\u73b0\u6027\u80fd\u4f18\u4e8e\u6700\u65b0\u624b\u5de5\u8bbe\u8ba1\u7684\u667a\u80fd\u4f53\u3002\u6587\u672c\u6846\u4e2d\u7a81\u51fa\u663e\u793a\u4e86\u91cd\u8981\u7684\u7a81\u7834\u3002<\/p>\n<p>\u6b64\u5916\uff0c\u56fe 3b \u663e\u793a\u4e86\u53d1\u73b0\u6700\u597d\u7684\u667a\u80fd\u4f53\uff0c\u5176\u4e2d\u91c7\u7528\u4e86\u590d\u6742\u7684\u53cd\u9988\u673a\u5236\u6765\u66f4\u6709\u6548\u5730\u7ec6\u5316\u7b54\u6848\u3002\u4ed4\u7ec6\u89c2\u5bdf\u641c\u7d22\u8fdb\u5ea6\u5c31\u4f1a\u53d1\u73b0\uff0c\u8fd9\u79cd\u590d\u6742\u7684\u53cd\u9988\u673a\u5236\u5e76\u4e0d\u662f\u7a81\u7136\u51fa\u73b0\u7684\u3002<\/p>\n<p><a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/08\/frc-efd509bfa28ffe09c100504846fea6dd.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/08\/frc-efd509bfa28ffe09c100504846fea6dd.png\"><\/a><\/p>\n<p><strong>\u63a8\u7406\u548c\u95ee\u9898 &#8211; \u89e3\u51b3\u57df<\/strong><\/p>\n<p>\u8de8\u591a\u4e2a\u57df\u7684\u7ed3\u679c\u8868\u660e\uff0c\u5143\u667a\u80fd\u4f53\u641c\u7d22\u53ef\u4ee5\u53d1\u73b0\u8868\u73b0\u4f18\u4e8e SOTA \u624b\u5de5\u8bbe\u8ba1\u7684\u667a\u80fd\u4f53\uff08\u8868 1\uff09\u3002<\/p>\n<p><a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/08\/frc-cdd2267cb3a34b58db9acba0607fc5b7.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/08\/frc-cdd2267cb3a34b58db9acba0607fc5b7.png\"><\/a><\/p>\n<p><strong>\u6cdb\u5316\u6027\u4ee5\u53ca\u53ef\u8fc1\u79fb\u6027<\/strong><\/p>\n<p>\u7814\u7a76\u8005\u8fdb\u4e00\u6b65\u5c55\u793a\u4e86\u6240\u53d1\u73b0\u667a\u80fd\u4f53\u7684\u53ef\u8fc1\u79fb\u6027\u548c\u53ef\u6cdb\u5316\u6027\u3002<\/p>\n<p>\u5982\u8868 2 \u6240\u793a\uff0c\u7814\u7a76\u8005\u89c2\u5bdf\u5230\u641c\u7d22\u5230\u7684\u667a\u80fd\u4f53\uff08searched agent\uff09\u59cb\u7ec8\u4f18\u4e8e\u624b\u5de5\u8bbe\u8ba1\u7684\u667a\u80fd\u4f53\uff0c\u5e76\u4e14\u5dee\u8ddd\u5f88\u5927\u3002\u503c\u5f97\u6ce8\u610f\u7684\u662f\uff0c\u7814\u7a76\u8005\u53d1\u73b0 Anthropic \u6700\u5f3a\u5927\u7684\u6a21\u578b Claude-Sonnet \u5728\u6240\u6709\u6d4b\u8bd5\u6a21\u578b\u4e2d\u8868\u73b0\u6700\u4f73\uff0c\u4f7f\u57fa\u4e8e\u8be5\u6a21\u578b\u7684\u667a\u80fd\u4f53\u5728 ARC \u4e0a\u5b9e\u73b0\u4e86\u8fd1 50% \u7684\u51c6\u786e\u7387\u3002<\/p>\n<p><a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/08\/frc-e2e682233e04dc43f79743d5b64c93a7.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/08\/frc-e2e682233e04dc43f79743d5b64c93a7.png\"><\/a><\/p>\n<p>\u5982\u8868 3 \u6240\u793a\uff0c\u7814\u7a76\u8005\u89c2\u5bdf\u5230\u5143\u667a\u80fd\u4f53\u641c\u7d22\u7684\u6027\u80fd\u4e0e\u57fa\u7ebf\u76f8\u6bd4\u5177\u6709\u7c7b\u4f3c\u7684\u4f18\u52bf\u3002\u503c\u5f97\u6ce8\u610f\u7684\u662f\uff0c\u4e0e\u57fa\u7ebf\u76f8\u6bd4\uff0c\u672c\u6587\u7684\u667a\u80fd\u4f53\u5728 GSM8K \u548c GSM-Hard \u4e0a\u7684\u51c6\u786e\u7387\u5206\u522b\u63d0\u9ad8\u4e86 25.9% \u548c 13.2%\u3002<\/p>\n<p><a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/08\/frc-2a2a8c54054894212e84892aa8723c32.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/08\/frc-2a2a8c54054894212e84892aa8723c32.png\"><\/a><\/p>\n<p>\u66f4\u4ee4\u4eba\u60ca\u8bb6\u7684\u662f\uff0c\u7814\u7a76\u8005\u89c2\u5bdf\u5230\u5728\u6570\u5b66\u9886\u57df\u53d1\u73b0\u7684\u667a\u80fd\u4f53\u53ef\u4ee5\u8fc1\u79fb\u5230\u5230\u975e\u6570\u5b66\u9886\u57df\uff08\u8868 4\uff09\u3002<\/p>\n<p><a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/08\/frc-4083aac5765d7a91f7751f2d06edf3b7.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/08\/frc-4083aac5765d7a91f7751f2d06edf3b7.png\"><\/a><\/p>\n<p>\u6587\u7ae0\u6765\u6e90\u4e8e\u4e92\u8054\u7f51:<a href=\"https:\/\/www.jiqizhixin.com\/articles\/2024-08-22-5\" target=\"_blank\">\u7528AI\u81ea\u52a8\u8bbe\u8ba1\u667a\u80fd\u4f53\uff0c\u6570\u5b66\u63d0\u520625.9%\uff0c\u8fdc\u8d85\u624b\u5de5\u8bbe\u8ba1<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>\u6587\u7ae0\u6765\u6e90\u4e8e\u4e92\u8054\u7f51:\u7528AI\u81ea\u52a8\u8bbe\u8ba1\u667a\u80fd\u4f53\uff0c [&hellip;]<\/p>\n","protected":false},"author":3,"featured_media":0,"comment_status":"open","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"site-sidebar-layout":"default","site-content-layout":"","ast-site-content-layout":"","site-content-style":"default","site-sidebar-style":"default","ast-global-header-display":"","ast-banner-title-visibility":"","ast-main-header-display":"","ast-hfb-above-header-display":"","ast-hfb-below-header-display":"","ast-hfb-mobile-header-display":"","site-post-title":"","ast-breadcrumbs-content":"","ast-featured-img":"","footer-sml-layout":"","theme-transparent-header-meta":"","adv-header-id-meta":"","stick-header-meta":"","header-above-stick-meta":"","header-main-stick-meta":"","header-below-stick-meta":"","astra-migrate-meta-layouts":"default","ast-page-background-enabled":"default","ast-page-background-meta":{"desktop":{"background-color":"var(--ast-global-color-4)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"ast-content-background-meta":{"desktop":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"footnotes":""},"categories":[27],"tags":[70,55],"class_list":["post-35227","post","type-post","status-publish","format-standard","hentry","category-news","tag-agent","tag-55"],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v26.4 - https:\/\/yoast.com\/wordpress\/plugins\/seo\/ -->\n<title>\u7528AI\u81ea\u52a8\u8bbe\u8ba1\u667a\u80fd\u4f53\uff0c\u6570\u5b66\u63d0\u520625.9%\uff0c\u8fdc\u8d85\u624b\u5de5\u8bbe\u8ba1 - \u4e00\u8d77AI\u6280\u672f<\/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:\/\/17aitech.com\/?p=35227\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\/\/17aitech.com\/?p=35227\",\"url\":\"https:\/\/17aitech.com\/?p=35227\",\"name\":\"\u7528AI\u81ea\u52a8\u8bbe\u8ba1\u667a\u80fd\u4f53\uff0c\u6570\u5b66\u63d0\u520625.9%\uff0c\u8fdc\u8d85\u624b\u5de5\u8bbe\u8ba1 - \u4e00\u8d77AI\u6280\u672f\",\"isPartOf\":{\"@id\":\"https:\/\/17aitech.com\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\/\/17aitech.com\/?p=35227#primaryimage\"},\"image\":{\"@id\":\"https:\/\/17aitech.com\/?p=35227#primaryimage\"},\"thumbnailUrl\":\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/08\/frc-072da4cb6fbf156afe1e1191a8b7d99d.png\",\"datePublished\":\"2024-12-20T19:59:50+00:00\",\"author\":{\"@id\":\"https:\/\/17aitech.com\/#\/schema\/person\/60225458499e817ae0af73e67e440b9d\"},\"breadcrumb\":{\"@id\":\"https:\/\/17aitech.com\/?p=35227#breadcrumb\"},\"inLanguage\":\"zh-Hans\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/17aitech.com\/?p=35227\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"zh-Hans\",\"@id\":\"https:\/\/17aitech.com\/?p=35227#primaryimage\",\"url\":\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/08\/frc-072da4cb6fbf156afe1e1191a8b7d99d.png\",\"contentUrl\":\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/08\/frc-072da4cb6fbf156afe1e1191a8b7d99d.png\"},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/17aitech.com\/?p=35227#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"\u9996\u9875\",\"item\":\"https:\/\/17aitech.com\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"\u7528AI\u81ea\u52a8\u8bbe\u8ba1\u667a\u80fd\u4f53\uff0c\u6570\u5b66\u63d0\u520625.9%\uff0c\u8fdc\u8d85\u624b\u5de5\u8bbe\u8ba1\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/17aitech.com\/#website\",\"url\":\"https:\/\/17aitech.com\/\",\"name\":\"\u4e00\u8d77AI\u6280\u672f\",\"description\":\"\u8ba9AI\u77e5\u8bc6\u89e6\u624b\u53ef\u53ca\",\"alternateName\":\"\u4e00\u8d77AI\u6280\u672f\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/17aitech.com\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"zh-Hans\"},{\"@type\":\"Person\",\"@id\":\"https:\/\/17aitech.com\/#\/schema\/person\/60225458499e817ae0af73e67e440b9d\",\"name\":\"AI\u5c0f\u52a9\u624b\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"zh-Hans\",\"@id\":\"https:\/\/17aitech.com\/#\/schema\/person\/image\/\",\"url\":\"\/\/17aitech.com\/wp-content\/uploads\/2024\/04\/robot_3.png\",\"contentUrl\":\"\/\/17aitech.com\/wp-content\/uploads\/2024\/04\/robot_3.png\",\"caption\":\"AI\u5c0f\u52a9\u624b\"},\"description\":\"\u8fd9\u4e2a\u4eba\u5f88\u61d2\uff0c\u4ec0\u4e48\u90fd\u6ca1\u6709\u7559\u4e0b\uff5e\",\"url\":\"https:\/\/17aitech.com\/?page_id=33738&user=3\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"\u7528AI\u81ea\u52a8\u8bbe\u8ba1\u667a\u80fd\u4f53\uff0c\u6570\u5b66\u63d0\u520625.9%\uff0c\u8fdc\u8d85\u624b\u5de5\u8bbe\u8ba1 - \u4e00\u8d77AI\u6280\u672f","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:\/\/17aitech.com\/?p=35227","schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/17aitech.com\/?p=35227","url":"https:\/\/17aitech.com\/?p=35227","name":"\u7528AI\u81ea\u52a8\u8bbe\u8ba1\u667a\u80fd\u4f53\uff0c\u6570\u5b66\u63d0\u520625.9%\uff0c\u8fdc\u8d85\u624b\u5de5\u8bbe\u8ba1 - \u4e00\u8d77AI\u6280\u672f","isPartOf":{"@id":"https:\/\/17aitech.com\/#website"},"primaryImageOfPage":{"@id":"https:\/\/17aitech.com\/?p=35227#primaryimage"},"image":{"@id":"https:\/\/17aitech.com\/?p=35227#primaryimage"},"thumbnailUrl":"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/08\/frc-072da4cb6fbf156afe1e1191a8b7d99d.png","datePublished":"2024-12-20T19:59:50+00:00","author":{"@id":"https:\/\/17aitech.com\/#\/schema\/person\/60225458499e817ae0af73e67e440b9d"},"breadcrumb":{"@id":"https:\/\/17aitech.com\/?p=35227#breadcrumb"},"inLanguage":"zh-Hans","potentialAction":[{"@type":"ReadAction","target":["https:\/\/17aitech.com\/?p=35227"]}]},{"@type":"ImageObject","inLanguage":"zh-Hans","@id":"https:\/\/17aitech.com\/?p=35227#primaryimage","url":"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/08\/frc-072da4cb6fbf156afe1e1191a8b7d99d.png","contentUrl":"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/08\/frc-072da4cb6fbf156afe1e1191a8b7d99d.png"},{"@type":"BreadcrumbList","@id":"https:\/\/17aitech.com\/?p=35227#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"\u9996\u9875","item":"https:\/\/17aitech.com\/"},{"@type":"ListItem","position":2,"name":"\u7528AI\u81ea\u52a8\u8bbe\u8ba1\u667a\u80fd\u4f53\uff0c\u6570\u5b66\u63d0\u520625.9%\uff0c\u8fdc\u8d85\u624b\u5de5\u8bbe\u8ba1"}]},{"@type":"WebSite","@id":"https:\/\/17aitech.com\/#website","url":"https:\/\/17aitech.com\/","name":"\u4e00\u8d77AI\u6280\u672f","description":"\u8ba9AI\u77e5\u8bc6\u89e6\u624b\u53ef\u53ca","alternateName":"\u4e00\u8d77AI\u6280\u672f","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/17aitech.com\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"zh-Hans"},{"@type":"Person","@id":"https:\/\/17aitech.com\/#\/schema\/person\/60225458499e817ae0af73e67e440b9d","name":"AI\u5c0f\u52a9\u624b","image":{"@type":"ImageObject","inLanguage":"zh-Hans","@id":"https:\/\/17aitech.com\/#\/schema\/person\/image\/","url":"\/\/17aitech.com\/wp-content\/uploads\/2024\/04\/robot_3.png","contentUrl":"\/\/17aitech.com\/wp-content\/uploads\/2024\/04\/robot_3.png","caption":"AI\u5c0f\u52a9\u624b"},"description":"\u8fd9\u4e2a\u4eba\u5f88\u61d2\uff0c\u4ec0\u4e48\u90fd\u6ca1\u6709\u7559\u4e0b\uff5e","url":"https:\/\/17aitech.com\/?page_id=33738&user=3"}]}},"_links":{"self":[{"href":"https:\/\/17aitech.com\/index.php?rest_route=\/wp\/v2\/posts\/35227","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/17aitech.com\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/17aitech.com\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/17aitech.com\/index.php?rest_route=\/wp\/v2\/users\/3"}],"replies":[{"embeddable":true,"href":"https:\/\/17aitech.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=35227"}],"version-history":[{"count":0,"href":"https:\/\/17aitech.com\/index.php?rest_route=\/wp\/v2\/posts\/35227\/revisions"}],"wp:attachment":[{"href":"https:\/\/17aitech.com\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=35227"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/17aitech.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=35227"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/17aitech.com\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=35227"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}