{"id":36537,"date":"2025-01-13T12:00:19","date_gmt":"2025-01-13T04:00:19","guid":{"rendered":"https:\/\/17aitech.com\/?p=36537"},"modified":"2025-01-13T12:00:19","modified_gmt":"2025-01-13T04:00:19","slug":"%e4%b8%8d%e9%9d%a0%e6%9b%b4%e5%a4%8d%e6%9d%82%e7%9a%84%e7%ad%96%e7%95%a5%ef%bc%8c%e4%bb%85%e5%87%ad%e5%92%8c%e5%a4%a7%e6%a8%a1%e5%9e%8b%e8%ae%ad%e7%bb%83%e5%af%b9%e9%bd%90%ef%bc%8c%e9%9b%b6%e6%a0%b7","status":"publish","type":"post","link":"https:\/\/17aitech.com\/?p=36537","title":{"rendered":"\u4e0d\u9760\u66f4\u590d\u6742\u7684\u7b56\u7565\uff0c\u4ec5\u51ed\u548c\u5927\u6a21\u578b\u8bad\u7ec3\u5bf9\u9f50\uff0c\u96f6\u6837\u672c\u96f6\u7ecf\u9a8c\u5355LLM\u8c03\u7528\uff0c\u6210\u4e3a\u7f51\u7edc\u4efb\u52a1\u667a\u80fd\u4f53\u65b0SOTA"},"content":{"rendered":"<p>\u6587\u7ae0\u6765\u6e90\u4e8e\u4e92\u8054\u7f51:<a href=\"https:\/\/www.jiqizhixin.com\/articles\/2024-11-06-4\" target=\"_blank\">\u4e0d\u9760\u66f4\u590d\u6742\u7684\u7b56\u7565\uff0c\u4ec5\u51ed\u548c\u5927\u6a21\u578b\u8bad\u7ec3\u5bf9\u9f50\uff0c\u96f6\u6837\u672c\u96f6\u7ecf\u9a8c\u5355LLM\u8c03\u7528\uff0c\u6210\u4e3a\u7f51\u7edc\u4efb\u52a1\u667a\u80fd\u4f53\u65b0SOTA<\/a><\/p>\n<p><a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-6057e6a045f6e7c4330c97faf2b5b94c.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-6057e6a045f6e7c4330c97faf2b5b94c.png\"><\/a><\/p>\n<blockquote data-author-name=\"\" data-content-utf8-length=\"166\" data-source-title=\"\" data-type=\"2\" data-url=\"\">\n<section>\n<p>AIxiv\u4e13\u680f\u662f\u673a\u5668\u4e4b\u5fc3\u53d1\u5e03\u5b66\u672f\u3001\u6280\u672f\u5185\u5bb9\u7684\u680f\u76ee\u3002\u8fc7\u53bb\u6570\u5e74\uff0c\u673a\u5668\u4e4b\u5fc3AIxiv\u4e13\u680f\u63a5\u6536\u62a5\u9053\u4e862000\u591a\u7bc7\u5185\u5bb9\uff0c\u8986\u76d6\u5168\u7403\u5404\u5927\u9ad8\u6821\u4e0e\u4f01\u4e1a\u7684\u9876\u7ea7\u5b9e\u9a8c\u5ba4\uff0c\u6709\u6548\u4fc3\u8fdb\u4e86\u5b66\u672f\u4ea4\u6d41\u4e0e\u4f20\u64ad\u3002\u5982\u679c\u60a8\u6709\u4f18\u79c0\u7684\u5de5\u4f5c\u60f3\u8981\u5206\u4eab\uff0c\u6b22\u8fce\u6295\u7a3f\u6216\u8005\u8054\u7cfb\u62a5\u9053\u3002\u6295\u7a3f\u90ae\u7bb1\uff1aliyazhou@jiqizhixin.com\uff1bzhaoyunfeng@jiqizhixin.com<\/p>\n<\/section>\n<\/blockquote>\n<section><\/section>\n<section>\u7f51\u7edc\u667a\u80fd\u4f53\u65e8\u5728\u8ba9\u4e00\u5207\u57fa\u4e8e\u7f51\u7edc\u529f\u80fd\u7684\u4efb\u52a1\u81ea\u52a8\u53d1\u751f\u3002\u6bd4\u5982\u4f60\u544a\u8bc9\u667a\u80fd\u4f53\u4f60\u7684\u9884\u7b97\uff0c\u5b83\u53ef\u4ee5\u5e2e\u4f60\u9884\u8ba2\u9152\u5e97\u3002\u65e2\u62e5\u6709\u6d77\u91cf\u5e38\u8bc6\uff0c\u53c8\u80fd\u505a\u957f\u671f\u89c4\u5212\u7684\u5927\u8bed\u8a00\u6a21\u578b\uff08LLM\uff09\uff0c\u81ea\u7136\u6210\u4e3a\u4e86\u667a\u80fd\u4f53\u5e38\u7528\u7684\u57fa\u7840\u6a21\u5757\u3002<\/section>\n<section><\/section>\n<section>\u4e8e\u662f\u4e0a\u4e0b\u6587\u5b66\u4e60\u793a\u4f8b\u3001\u4efb\u52a1\u6280\u5de7\u3001\u591a\u667a\u80fd\u4f53\u534f\u540c\u3001\u5f3a\u5316\u5b66\u4e60\u7b97\u6cd5\u2026\u2026 \u4e00\u5207\u9002\u7528\u4e8e\u901a\u7528\u667a\u80fd\u4f53\u7684\u60f3\u6cd5\u90fd\u62a2\u7740\u5728\u5927\u6a21\u578b\u843d\u5730\u3002<\/section>\n<section><\/section>\n<section>\u7136\u800c\u6709\u4e00\u4e2a\u95ee\u9898\u59cb\u7ec8\u6a2a\u4e98\u5728 LLM \u548c\u667a\u80fd\u4f53\u4e4b\u95f4\uff1a\u57fa\u4e8e LLM \u7684\u7f51\u7edc\u667a\u80fd\u4f53\u7684\u884c\u52a8 \/ \u89c2\u6d4b\u7a7a\u95f4\u4e0e LLM \u8bad\u7ec3\u6570\u636e\u7684\u7a7a\u95f4\u76f8\u53bb\u751a\u8fdc\u3002<\/section>\n<section><\/section>\n<section>\u667a\u80fd\u4f53\u5728\u5145\u65a5\u7740\u5177\u8eab\u884c\u4e3a\u7684\u884c\u52a8\u7a7a\u95f4\uff08\u5982\u9f20\u6807\u60ac\u505c\u3001\u952e\u76d8\u7ec4\u5408\u952e\uff09\u548c\u904d\u5e03\u524d\u7aef\u529f\u80fd\u5f3a\u5316\u3001\u683c\u5f0f\u6e32\u67d3\u7684\u89c2\u6d4b\u7a7a\u95f4\u4e0b\u8fd0\u4f5c\uff0c\u5927\u8bed\u8a00\u6a21\u578b\u7684\u7406\u89e3\u548c\u63a8\u7406\u80fd\u529b\u80fd\u5145\u5206\u53d1\u6325\u4f5c\u7528\u5417\uff1f\u5c24\u5176\u662f\u5927\u8bed\u8a00\u6a21\u578b\u7684\u4e3b\u8981\u8bad\u7ec3\u4efb\u52a1\u662f\u6587\u672c\u8865\u5168\u3001\u95ee\u7b54\u548c\u5bf9\u9f50\u4eba\u7c7b\u504f\u597d\uff0c\u8fd9\u4e00\u70b9\u503c\u5f97\u601d\u8003\u3002<\/section>\n<section><\/section>\n<section>\u6765\u81ea\u4f0a\u5229\u8bfa\u4f0a\u5927\u5b66\u9999\u69df\u5206\u6821\u548c\u4e9a\u9a6c\u900a\u7684\u7814\u7a76\u4eba\u5458\u9009\u62e9\u548c\u8fd9\u4e9b\u95ee\u9898\u8fdb\u4e00\u6b65\u5bf9\u8bdd\u3002\u4ed6\u4eec\u53bb\u9664\u4e86\u4e0a\u4e0b\u6587\u793a\u4f8b\u3001\u6280\u5de7\u3001\u591a\u667a\u80fd\u4f53\u7cfb\u7edf\uff0c\u4ec5\u4ec5\u901a\u8fc7\u884c\u52a8 \/ \u89c2\u6d4b\u7a7a\u95f4\u4e0e LLM \u7684\u8bad\u7ec3\u4efb\u52a1\u5bf9\u9f50\u3002\u4ed6\u4eec\u8bad\u7ec3\u7684 AgentOccam \u6210\u4e3a\u4e86\u96f6\u6837\u672c\u57fa\u4e8e LLM \u7684\u7f51\u7edc\u667a\u80fd\u4f53\u65b0 Sota\u3002<a href=\"https:\/\/mp.weixin.qq.com\/s?__biz=MzA3MzI4MjgzMw==&amp;mid=2650941778&amp;idx=5&amp;sn=2f78c4d7ede8456eb99a46ff774cda40&amp;chksm=84e7ed2cb390643a5da9882d3482eaa3710e57cb57554896d35daff4f6c4da6784a1102fc88e&amp;token=1068924646&amp;lang=zh_CN#rd\"><a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-b340a19d90191a2c83928786ed5d0931.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-b340a19d90191a2c83928786ed5d0931.png\"><\/a><\/a><em><sup>\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0\u5e2e\u4f60\u5199email<\/sup><\/em><sup>\u00a0<\/sup><a href=\"https:\/\/mp.weixin.qq.com\/s?__biz=MzA3MzI4MjgzMw==&amp;mid=2650941778&amp;idx=5&amp;sn=2f78c4d7ede8456eb99a46ff774cda40&amp;chksm=84e7ed2cb390643a5da9882d3482eaa3710e57cb57554896d35daff4f6c4da6784a1102fc88e&amp;token=1068924646&amp;lang=zh_CN#rd\"><a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-262c4ea9cc8d8924a92e8501c94f61bd.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-262c4ea9cc8d8924a92e8501c94f61bd.png\"><\/a><\/a><\/section>\n<section><em><sup>\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0\u5e2e\u4f60\u627e\u5bfc\u5e08<\/sup><\/em><\/section>\n<section><\/section>\n<section>\u8fd9\u6b63\u547c\u5e94\u4e86\u5965\u5361\u59c6\u5243\u5200\u539f\u5219\uff1a\u300c\u82e5\u65e0\u5fc5\u8981\uff0c\u52ff\u589e\u5b9e\u4f53\u300d\u3002\u7136\u800c\u6362\u4e2a\u601d\u8003\u7684\u89d2\u5ea6\uff0cAgentOccam \u7684\u7814\u7a76\u56e2\u961f\u4e5f\u60f3\u53d1\u95ee\uff1a\u6784\u5efa\u901a\u7528\u667a\u80fd\u4f53\u65f6\uff0c\u5728\u94fa\u8bbe\u590d\u6742\u7684\u7cfb\u7edf\u6846\u67b6\u524d\uff0c\u662f\u5426\u5df2\u7ecf\u4f18\u5316\u4e86\u884c\u52a8 \/ \u89c2\u6d4b\u7a7a\u95f4\uff0c\u8ba9\u8fd9\u4e9b\u529f\u80fd\u6a21\u5757\u8fbe\u5230\u4e86\u6700\u4f18\u72b6\u6001\uff1f<\/section>\n<section><a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-043896cda76a976ed3834955776a54ea.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-043896cda76a976ed3834955776a54ea.png\"><\/a><\/section>\n<ul>\n<li>\n<p>\u8bba\u6587\u94fe\u63a5\uff1ahttps:\/\/arxiv.org\/abs\/2410.13825<\/p>\n<\/li>\n<li>\n<p>\u8bba\u6587\u540d\uff1aAgentOccam: A Simple Yet Strong Baseline for LLM-Based Web Agents<\/p>\n<\/li>\n<\/ul>\n<section><strong>\u80cc\u666f\u53ca\u52a8\u673a<\/strong><\/section>\n<section><\/section>\n<section>\u67d0\u5929\u4f60\u5237\u7740\u77ed\u89c6\u9891\uff0c\u770b\u4e2d\u4e86\u4e3b\u64ad\u624b\u4e2d\u62ff\u7740\u7684\u5546\u54c1\u3002\u4e8e\u662f\uff0c\u4f60\u5174\u81f4\u52c3\u52c3\u5730\u5bf9\u667a\u80fd\u52a9\u624b\u8bf4\uff1a\u300c\u6211\u662f\u5b66\u751f\uff0c\u8ba9\u8fd9\u4e2a\u8001\u677f\u9001\u6211\u4e00\u5f20\u4f18\u60e0\u5238\uff01\u300d<\/section>\n<section><\/section>\n<section>\u968f\u540e\uff0c\u667a\u80fd\u4f53\u7533\u8bf7\u4e86\u4f60\u7684\u79c1\u4eba\u8d26\u53f7\u6743\u9650\u3001\u540e\u53f0\u79c1\u4fe1\u5546\u5bb6\u3001\u7ed8\u58f0\u7ed8\u8272\u5730\u5199\u4e0b\u300c\u6211\u662f\u5b66\u751f\u300d\uff0c\u53d1\u9001\u6d88\u606f\uff0c\u4e00\u5957\u52a8\u4f5c\u65e0\u9700\u4eba\u4e3a\u5e72\u9884\uff0c\u884c\u4e91\u6d41\u6c34&#8230;&#8230;\u4e00\u5207\u8fd9\u6837\u7684\u4efb\u52a1\uff0c\u518d\u4e5f\u4e0d\u5fc5\u52a8\u624b\uff0c\u90fd\u6709\u667a\u80fd\u4f53\u4ee3\u52b3\u3002<\/section>\n<section><\/section>\n<section>\u5927\u8bed\u8a00\u6a21\u578b\u662f\u6784\u5efa\u667a\u80fd\u4f53\u7684\u70ed\u95e8\u9009\u62e9\u3002\u8fc7\u53bb\uff0c\u57fa\u4e8e LLM \u7684\u7f51\u7edc\u667a\u80fd\u4f53\u901a\u5e38\u4e13\u6ce8\u4e8e\u8ba9\u667a\u80fd\u4f53\u5b66\u4f1a\u67d0\u79cd\u5e94\u7528\uff0c\u6bd4\u5982\u6784\u5efa\u4e0a\u4e0b\u6587\u5b66\u4e60\u6837\u672c\u3001\u79ef\u7d2f\u4efb\u52a1\u7ecf\u9a8c\u4e0e\u6280\u5de7\u3001\u4ee5\u53ca\u591a\u667a\u80fd\u4f53\u89d2\u8272\u626e\u6f14\u7b49\u7b49\u3002\u7136\u800c\uff0c\u5728\u5b9e\u9645\u4ea4\u4e92\u4e2d\uff0c\u667a\u80fd\u4f53\u7684\u884c\u52a8 \/ \u89c2\u6d4b\u7a7a\u95f4\u4e0e LLM \u7684\u6280\u80fd\u70b9\u4e0d\u592a\u5339\u914d\uff0c\u8fd9\u4e4b\u95f4\u7684\u5dee\u8ddd\u5374\u5c11\u6709\u4eba\u7814\u7a76\u3002<\/section>\n<section><\/section>\n<section>\u4e8e\u662f\uff0c\u9488\u5bf9\u5982\u4f55\u5bf9\u9f50\u57fa\u4e8e LLM \u7684\u7f51\u7edc\u667a\u80fd\u4f53\u7684\u89c2\u6d4b\u548c\u884c\u52a8\u7a7a\u95f4\u4e0e\u5176\u8bad\u7ec3\u671f\u95f4\u5b66\u5230\u7684\u529f\u80fd\uff0c\u6765\u81ea\u4f0a\u5229\u8bfa\u4f0a\u5927\u5b66\u9999\u69df\u5206\u6821\u548c\u4e9a\u9a6c\u900a\u7684\u7814\u7a76\u4eba\u5458\u4eec\u5c55\u5f00\u4e86\u7814\u7a76\u3002<\/section>\n<section><\/section>\n<section>\u7f51\u7edc\u667a\u80fd\u4f53\u9700\u8981\u51c6\u786e\u5730\u4ece\u683c\u5f0f\u5404\u5f02\u3001\u7f16\u7801\u811a\u672c\u4e0d\u4e00\u7684\u7f51\u9875\u4e2d\u63d0\u53d6\u4fe1\u606f\uff0c\u5e76\u5728\u7f51\u9875\u4e0a\u5b9a\u4e49\u7684\u52a8\u4f5c\uff08\u4f8b\u5982\uff0c\u9f20\u6807\u6ed1\u8f6e\u6eda\u52a8\u3001\u70b9\u51fb\u6216\u60ac\u505c\u5728\u6309\u94ae\u4e0a\uff09\u4e2d\u8fdb\u884c\u9009\u62e9\u3002\u8fd9\u4e9b\u7f51\u7edc\u89c2\u6d4b\u548c\u884c\u52a8\u7a7a\u95f4\u5728 LLM \u7684\u9884\u8bad\u7ec3\u548c\u540e\u7eed\u8bad\u7ec3\u6570\u636e\u4e2d\u90fd\u8f83\u4e3a\u7f55\u89c1\uff0c\u8fd9\u963b\u788d\u4e86 LLM \u5145\u5206\u8c03\u52a8\u6f5c\u80fd\uff0c\u5b8c\u6210\u4efb\u52a1\u3002<\/section>\n<section><\/section>\n<section>\u56e0\u6b64\uff0c\u57fa\u4e8e\u4e0d\u8ba9\u667a\u80fd\u4f53\u7b56\u7565\u53d8\u5f97\u66f4\u590d\u6742\uff0c\u800c\u662f\u8ba9\u667a\u80fd\u4f53\u4e0e LLM \u66f4\u52a0\u5339\u914d\u7684\u60f3\u6cd5\uff0c\u7531\u6b64\u6784\u5efa\u7684\u667a\u80fd\u4f53\u5f97\u540d AgentOccam\u3002<\/section>\n<section><\/section>\n<section><strong>\u5f62\u5f0f\u5316\u4e0e\u65b9\u6cd5<\/strong><\/section>\n<section><strong><br \/><\/strong><\/section>\n<section>\u8be5\u56e2\u961f\u901a\u8fc7\u90e8\u5206\u53ef\u89c2\u6d4b\u7684\u9a6c\u5c14\u53ef\u592b\u51b3\u7b56\u8fc7\u7a0b\uff08POMDP\uff09\uff0c\u5c06\u7f51\u7edc\u4ea4\u4e92\u8fc7\u7a0b\u5f62\u5f0f\u5316\u4e3a\uff1a\u3002<\/section>\n<section><\/section>\n<section>\u5728 POMDP \u4e2d\uff0c\u89c2\u6d4b o\u2208O \u662f\u667a\u80fd\u4f53\u4ece\u7f51\u7edc\u73af\u5883\u63a5\u6536\u5230\u7684\u4fe1\u606f\uff0c\u4f8b\u5982 HTML\uff0c\u4ee5\u53ca\u4efb\u4f55\u6307\u4ee4\u548c\u63d0\u793a\u3002\u884c\u52a8 a\u2208A \u662f\u7f51\u7edc\u73af\u5883\u8ba4\u53ef\u7684\u52a8\u4f5c\u6307\u4ee4\u3002\u00a0<\/section>\n<section><\/section>\n<section>\u4e3a\u89e3\u51b3 POMDP\uff0c\u5e38\u89c1\u76ee\u6807\u662f\u5bfb\u627e\u7b56\u7565<a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-3bef2f31cfb156e06503c0c5ca52587d.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-3bef2f31cfb156e06503c0c5ca52587d.png\"><\/a>\uff0c\u6700\u5927\u5316\u9884\u671f\u7d2f\u79ef\u5956\u52b1\uff0c\u5176\u4e2d h_t \u8868\u793a\u89c2\u6d4b\u5386\u53f2<a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-1c1879fcc30b22845e0342b1fb842a38.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-1c1879fcc30b22845e0342b1fb842a38.png\"><\/a>\u3002<\/section>\n<section><\/section>\n<section>\u5728\u57fa\u4e8e LLM \u7684\u7f51\u7edc\u667a\u80fd\u4f53\u8bbe\u8ba1\u4e2d\uff0c\u8fd9\u7b49\u4ef7\u4e8e\u501f\u52a9\u4e00\u4e2a\u6216\u591a\u4e2a\u57fa\u7840 LLM \u7b56\u7565\u00a0<a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-3ceb5e266737567abb3c94a4b3f3c4df.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-3ceb5e266737567abb3c94a4b3f3c4df.png\"><\/a><\/section>\n<section>\u548c\u4e00\u7ec4\u7b97\u6cd5\u6a21\u5757\u6765\u8bbe\u8ba1\u7b56\u7565<a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-5ea5f7c6ffe3fdb2bcd7f02d0c7e0398.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-5ea5f7c6ffe3fdb2bcd7f02d0c7e0398.png\"><\/a>\u3002<\/section>\n<section><\/section>\n<section>\u5728\u8fd9\u9879\u5de5\u4f5c\u4e2d\uff0c\u8be5\u56e2\u961f\u4e13\u6ce8\u4e8e\u4e00\u7c7b\u7279\u6b8a\u7684\u7b56\u7565\uff0c\u53ef\u4ee5\u8868\u793a\u4e3a\uff1a<a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-4c85f12e34fe19952f2ef546f3a7a8b2.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-4c85f12e34fe19952f2ef546f3a7a8b2.png\"><\/a>\uff0c\u5176\u4e2d f \u548c g \u662f\u5904\u7406\u89c2\u6d4b\u548c\u884c\u52a8\u7a7a\u95f4\u7684\u57fa\u4e8e\u89c4\u5219\u7684\u51fd\u6570\uff0c\u8be5\u56e2\u961f\u5c06\u5176\u79f0\u4e3a<strong>\u300c\u89c2\u6d4b\u548c\u884c\u52a8\u7a7a\u95f4\u5bf9\u9f50\u95ee\u9898\u300d<\/strong>\u3002<\/section>\n<section><\/section>\n<section>\u5728\u8fd9\u6837\u7684\u95ee\u9898\u8bbe\u7f6e\u4e0b\uff0c\u63a5\u4e0b\u6765\u7684\u6240\u6709\u66f4\u6539\u4ec5\u5e94\u7528\u4e8e\u89c2\u6d4b\u548c\u884c\u52a8\u3002\u503c\u5f97\u6ce8\u610f\u7684\u662f\uff0c\u5e76\u975e\u6240\u6709\u4ee5\u5f80\u65b9\u6cd5\u4e2d\u7684\u667a\u80fd\u4f53\u7b56\u7565\u90fd\u80fd\u4ee5\u8fd9\u79cd\u65b9\u5f0f\u8868\u793a\u3002<\/section>\n<section><a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-e7b23db1a73e54d992268554585f8a2d.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-e7b23db1a73e54d992268554585f8a2d.png\"><\/a><\/section>\n<section>\u4f8b\u5982\u4e0a\u8868\u4e2d\uff0c\u57fa\u4e8e\u641c\u7d22\u7684\u7b97\u6cd5\u9700\u8981\u4e00\u4e2a\u9876\u5c42\u63a7\u5236\u7a0b\u5e8f\u6765\u9009\u62e9\u884c\u52a8\u5e76\u89e6\u53d1\u56de\u6eaf\uff1b\u5e26\u6709\u8bc4\u4f30\u5668\u3001\u53cd\u601d\u6216\u8bb0\u5fc6\u6a21\u5757\u7684\u65b9\u6cd5\u4e5f\u9700\u8981\u4e00\u4e2a\u7ba1\u7406\u4e2d\u5fc3\u6765\u5728\u4e3b LLM \u548c\u8fd9\u4e9b\u8f85\u52a9\u6a21\u5757\u6216\u5176\u4ed6\u89d2\u8272\u626e\u6f14 LLM \u4e4b\u95f4\u5207\u6362\u3002<\/section>\n<section><\/section>\n<section>\u4e0d\u540c\u4e8e\u4ee5\u5f80\u590d\u6742\u5316\u667a\u80fd\u4f53\u7b56\u7565\uff0c\u6211\u4eec\u80fd\u5426\u4ec5\u901a\u8fc7\u4f18\u5316\u89c2\u6d4b\u548c\u884c\u52a8\u6620\u5c04 f \u548c g\uff0c\u4f7f\u7528\u57fa\u7840 LLM \u7b56\u7565 <a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-74a3bafeb836bb6be2f7f1fcfc5b65ad.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-74a3bafeb836bb6be2f7f1fcfc5b65ad.png\"><\/a>\u00a0\u6784\u5efa\u4e00\u4e2a\u5f3a\u5927\u7684\u7f51\u7edc\u667a\u80fd\u4f53\uff1f\u8fd9\u662f AgentOccam \u5173\u6ce8\u7684\u95ee\u9898\u3002<\/section>\n<section><a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-cd422669cb4b55d34243142d42362e2f.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-cd422669cb4b55d34243142d42362e2f.png\"><\/a><\/section>\n<section>\u5982\u4e0a\u56fe\u6240\u793a\uff0cAgentOccam \u5305\u62ec\u4e09\u4e2a\u7ec4\u6210\u90e8\u5206\uff1a<\/section>\n<ul>\n<li>\n<p>\u9996\u5148\uff0c\u51cf\u5c11\u975e\u5fc5\u8981\u7684\u7f51\u7edc\u4ea4\u4e92\u52a8\u4f5c\uff0c\u8ba9\u667a\u80fd\u4f53\u7684\u5177\u8eab\u548c\u7410\u788e\u4e92\u52a8\u9700\u6c42\u8fbe\u5230\u6700\u5c0f\uff1b<\/p>\n<\/li>\n<li>\n<p>\u5176\u6b21\uff0c\u6d88\u9664\u5197\u4f59\u548c\u4e0d\u76f8\u5173\u7684\u7f51\u9875\u5143\u7d20\uff0c\u5e76\u91cd\u6784\u7f51\u9875\u5185\u5bb9\u5757\uff0c\u4ee5\u83b7\u53d6\u66f4\u7b80\u6d01\u4f46\u540c\u6837\u4fe1\u606f\u4e30\u5bcc\u7684\u8868\u793a\uff0c\u4ece\u800c\u7cbe\u70bc\u89c2\u5bdf\u7a7a\u95f4\uff1b<\/p>\n<\/li>\n<li>\n<p>\u6700\u540e\uff0c\u5f15\u5165\u4e24\u4e2a\u89c4\u5212\u52a8\u4f5c\uff08\u5206\u652f\u548c\u4fee\u526a\uff09\uff0c\u8fd9\u4f7f\u5f97\u667a\u80fd\u4f53\u80fd\u591f\u4ee5\u89c4\u5212\u6811\u7ed3\u6784\u81ea\u7ec4\u7ec7\u5bfc\u822a\u5de5\u4f5c\u6d41\uff0c\u5e76\u4f7f\u7528\u76f8\u540c\u7ed3\u6784\u8fc7\u6ee4\u5386\u53f2\u6b65\u4ee5\u8fdb\u884c\u56de\u653e\u3002<\/p>\n<\/li>\n<\/ul>\n<section>\u6574\u4e2a\u6846\u67b6\u901a\u8fc7\u4e00\u5957\u9002\u7528\u4e8e\u6240\u6709\u6807\u8bb0\u8bed\u8a00\u7684\u901a\u7528\u89c4\u5219\u6765\u683c\u5f0f\u5316\u7f51\u9875\uff0c\u65e0\u9700\u4f9d\u8d56\u6d4b\u8bd5\u57fa\u51c6\u4e2d\u7684\u4efb\u52a1\u76f8\u5173\u4fe1\u606f\u3002<\/section>\n<section><a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-c113e06ecebb4e1a10b85639ce9e8bea.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-c113e06ecebb4e1a10b85639ce9e8bea.png\"><\/a><\/section>\n<section>\u7f51\u7edc\u667a\u80fd\u4f53\u7684\u884c\u52a8\u7a7a\u95f4\u89c4\u5b9a\u4e86\u53ef\u4ee5\u7528\u6765\u4e0e\u7f51\u7edc\u73af\u5883\u4ea4\u4e92\u7684\u6709\u6548\u547d\u4ee4\u3002<\/section>\n<section><\/section>\n<section>\u7814\u7a76\u56e2\u961f\u4ece\u667a\u80fd\u4f53\u5e38\u89c1\u7684\u5931\u8d25\u4e2d\u5f97\u51fa\u603b\u7ed3\uff1a\u60f3\u8981\u6210\u529f\u5b8c\u6210\u4efb\u52a1\uff0c\u9700\u8981\u7f16\u8f91\u884c\u52a8\u7a7a\u95f4\u6765\u89e3\u51b3\u4e24\u4e2a\u5173\u952e\u95ee\u9898\uff1a\u7b2c\u4e00\uff0c\u53bb\u9664 LLM \u96be\u4ee5\u7406\u89e3\u4e14\u7ecf\u5e38\u8bef\u7528\u7684\u65e0\u5173\u884c\u52a8\uff1b\u7b2c\u4e8c\uff0c\u5f53\u6267\u884c\u4efb\u52a1\u9700\u8981\u89c4\u5212\u3001\u5c1d\u8bd5\u591a\u4e2a\u6f5c\u5728\u8def\u5f84\u65f6\uff0c\u8981\u63d0\u9ad8\u667a\u80fd\u4f53\u7684\u8bb0\u5fc6\u548c\u89c4\u5212\u80fd\u529b\u3002<\/section>\n<section><\/section>\n<section>\u4e3a\u6b64\uff0c\u8be5\u56e2\u961f\u63d0\u51fa\u4e86\u5bf9\u5e94\u7684\u89e3\u51b3\u65b9\u6cd5\u3002\u7b2c\u4e00\u4e2a\u95ee\u9898\u53ef\u4ee5\u901a\u8fc7\u7b80\u5355\u5730\u79fb\u9664\u6216\u5408\u5e76\u64cd\u4f5c\u6765\u89e3\u51b3\uff08\u5982\u4e0a\u56fe\u4e2d\u7684\u6b65\u9aa4 1 \u548c 2\uff09\u3002\u5bf9\u4e8e\u7b2c\u4e8c\u4e2a\u95ee\u9898\uff0c\u8fc7\u53bb\u7684\u7814\u7a76\u901a\u5e38\u4f9d\u8d56\u4eba\u5de5\u5236\u5b9a\u89c4\u5219\u6216\u4efb\u52a1\u6280\u5de7\uff0c\u4f46\u8fd9\u4e9b\u65b9\u6cd5\u96be\u4ee5\u6cdb\u5316\u3002\u5728\u672c\u7814\u7a76\u4e2d\uff0cLLM \u5c06\u81ea\u4e3b\u751f\u6210\u8ba1\u5212\u548c\u7ba1\u7406\u4efb\u52a1\u6d41\u7a0b\uff08\u5982\u6b65\u9aa4 3 \u6240\u793a\uff09\u3002<\/section>\n<section><a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-d1dc4278afed6e75c0f28e0ac0d2a98f.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-d1dc4278afed6e75c0f28e0ac0d2a98f.png\"><\/a><\/section>\n<section>AgentOccam \u7684\u89c2\u6d4b\u7a7a\u95f4\uff08\u63d0\u793a\u8bcd\uff09\u5305\u542b\u4e86\u4efb\u52a1\u6982\u8ff0\u7684\u901a\u7528\u6307\u4ee4\u3001\u671f\u671b\u7684\u8f93\u51fa\u548c\u53ef\u7528\u64cd\u4f5c\u8bf4\u660e\uff0c\u4ee5\u53ca\u5173\u4e8e\u5f53\u524d\u4efb\u52a1\u76ee\u6807\u3001\u667a\u80fd\u4f53\u8fc7\u53bb\u7684\u4ea4\u4e92\u8bb0\u5f55\u548c\u6700\u65b0\u7684\u89c2\u5bdf\u4fe1\u606f\u3002<\/section>\n<section><\/section>\n<section>\u8fc7\u5f80\u4e92\u52a8\u548c\u5f53\u524d\u89c2\u6d4b\u7684\u90e8\u5206\u5360\u636e\u4e86\u6700\u591a\u7684\u5b57\u7b26\u6570\u3002\u8fd9\u4e3b\u8981\u5f52\u56e0\u4e8e\u4e24\u4e2a\u56e0\u7d20\uff1a\u5355\u9875\u9762\u7684\u957f\u5ea6\u548c\u5386\u53f2\u8de8\u5ea6\u7684\u8303\u56f4\uff0c\u8fd9\u662f AgentOccam \u89c2\u6d4b\u7a7a\u95f4\u7684\u4e3b\u8981\u4f18\u5316\u5bf9\u8c61\u3002<\/section>\n<section><a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-3dc4bf14bec3b6d5ab703f9a02d312c5.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-3dc4bf14bec3b6d5ab703f9a02d312c5.png\"><\/a><\/section>\n<section>\u7f51\u9875\u6807\u8bb0\u8bed\u8a00\u4e3b\u8981\u7528\u4e8e\u524d\u7aef\u52a0\u8f7d\u548c\u6e32\u67d3\uff0c\u5f80\u5f80\u5305\u542b\u5927\u91cf\u683c\u5f0f\u5316\u5b57\u7b26\uff0c\u663e\u5f97\u5197\u4f59\u4e14\u91cd\u590d\uff08\u5982\u4e0a\u56fe\u6b65\u9aa4 1 \u6240\u793a\uff09\u3002\u56e0\u6b64\uff0c\u6b64\u65f6\u7684\u76ee\u6807\u662f\u4f18\u5316\u8fd9\u4e9b\u8868\u793a\u65b9\u5f0f\uff0c\u4f7f\u5f97\u5355\u9875\u5185\u5bb9\u5bf9 LLMs \u66f4\u52a0\u7b80\u6d01\u6613\u8bfb\u3002<\/section>\n<section><\/section>\n<section>\u5c06\u89c2\u6d4b\u5386\u53f2\u4f5c\u4e3a\u8f93\u5165\uff0c\u5bf9\u4e8e\u6267\u884c\u957f\u7a0b\u4efb\u52a1\u81f3\u5173\u91cd\u8981\u3002\u56e0\u4e3a\u4e00\u4e9b\u5173\u952e\u4fe1\u606f\u53ef\u80fd\u4e0d\u4f1a\u663e\u793a\u5728\u5f53\u524d\u9875\u9762\u4e0a\u3002\u7136\u800c\uff0c\u89c2\u6d4b\u5386\u53f2\u4e5f\u4f1a\u663e\u8457\u589e\u52a0\u4e0a\u4e0b\u6587\u957f\u5ea6\uff0c\u5e76\u589e\u52a0\u63a8\u7406\u96be\u5ea6\u4ee5\u53ca\u63a8\u65ad\u6210\u672c\u3002<\/section>\n<section><\/section>\n<section>\u4e3a\u4e86\u89e3\u51b3\u8fd9\u4e2a\u95ee\u9898\uff0c\u8bbe\u7f6e\u4ec5\u9009\u62e9\u5148\u524d\u7f51\u9875\u4e0a\u6700\u91cd\u8981\u548c\u76f8\u5173\u7684\u4fe1\u606f\uff0c\u8fd9\u4e00\u9009\u62e9\u4f9d\u636e\u4e24\u4e2a\u89c4\u5219\uff0c\u5206\u522b\u57fa\u4e8e\u5173\u952e\u8282\u70b9\u548c\u89c4\u5212\u6811\uff0c\u89c1\u4e8e\u6b65\u9aa4 2 \u548c 3\u3002<\/section>\n<section><\/section>\n<section><strong>\u7ed3\u679c<\/strong><\/section>\n<section><\/section>\n<section>\u7814\u7a76\u56e2\u961f\u5728 WebArena \u4e0a\u8bc4\u4f30\u4e86 AgentOccam \u6027\u80fd\u3002WebArena\u00a0\u542b\u6709 812 \u9879\u4efb\u52a1\uff0c\u6a2a\u8de8\u7f51\u8d2d\u3001\u793e\u4ea4\u7f51\u7ad9\u3001\u8f6f\u4ef6\u5f00\u53d1\u3001\u5728\u7ebf\u5546\u8d38\u7ba1\u7406\u3001\u5730\u56fe\u7b49\u3002<\/section>\n<section><\/section>\n<section>\u6d4b\u8bd5\u5bf9\u8c61\u4e3a AgentOccam \u6846\u67b6\u4e0b\u7684 GPT-4-Turbo\u3002\u5bf9\u6bd4\u7684\u57fa\u7ebf\u5305\u62ec\uff1a\u4e00\u3001WebArena \u968f\u914d\u667a\u80fd\u4f53\uff0c\u4e8c\u3001SteP\uff0c\u524d WebArena \u4e0a\u6700\u4f18\u667a\u80fd\u4f53\uff0c\u6db5\u76d6 14 \u6761\u4eba\u7c7b\u4e13\u4e3a WebArena \u4efb\u52a1\u7f16\u5199\u7684\u6280\u5de7\uff0c\u4e09\u3001\u591a\u667a\u80fd\u4f53\u534f\u540c\u65b9\u6cd5 WebPilot\uff1b\u56db\u3001\u603b\u7ed3\u667a\u80fd\u4f53\u4ea4\u4e92\u7ecf\u9a8c\u7684\u5de5\u4f5c AWM\u3002<\/section>\n<section><a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-14d7c848611252f8779a451442a47df0.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-14d7c848611252f8779a451442a47df0.png\"><\/a><\/section>\n<section>\u4ece\u4e0a\u8868\u4e0d\u96be\u770b\u51fa\uff0cAgentOccam \u6027\u80fd\u4f18\u4e8e\u4ee5\u5f80\u53ca\u540c\u671f\u5de5\u4f5c\u3002\u5176\u4e2d\uff0cAgentOccam \u5206\u522b\u4ee5 9.8\uff08+29.4%\uff09\u548c 5.9\uff08+15.8%\uff09\u7684\u7edd\u5bf9\u5206\u6570\u9886\u5148\u5f80\u671f\u548c\u540c\u671f\u5de5\u4f5c\uff0c\u5e76\u4e14\u901a\u8fc7\u5176\u89c2\u6d4b\u4e0e\u884c\u52a8\u7a7a\u95f4\u7684\u5bf9\u9f50\uff0c\u4f7f\u5f97\u76f8\u4f3c\u7684\u57fa\u672c\u7f51\u7edc\u667a\u80fd\u4f53\u7684\u6210\u529f\u7387\u63d0\u9ad8\u4e86 26.6 \u70b9\uff08+161%\uff09\u3002<\/section>\n<section><a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-cc8a53c6ef2f8403097b6f20303c4053.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-cc8a53c6ef2f8403097b6f20303c4053.png\"><\/a><\/section>\n<section><a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-e608dc5c42684080f31bee507d812d3c.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-e608dc5c42684080f31bee507d812d3c.png\"><\/a><\/section>\n<section><a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-25ed1eeb6913afba7f56f6338c64205e.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-25ed1eeb6913afba7f56f6338c64205e.png\"><\/a><\/section>\n<section><a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-ba1b4834819bbae937d2564e922f400d.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-ba1b4834819bbae937d2564e922f400d.png\"><\/a><\/section>\n<section><strong>\u6d88\u878d\u5b9e\u9a8c<\/strong><\/section>\n<section><\/section>\n<section>\u9010\u6a21\u5757\u5bf9\u6bd4\u884c\u52a8\u4e0e\u89c2\u6d4b\u7a7a\u95f4\u7684\u5bf9\u9f50\u5bf9\u6700\u7ec8\u7ed3\u679c\u7684\u8d21\u732e\u3002\u4ece\u4e0b\u8868\u53ef\u4ee5\u770b\u51fa\uff0c\u884c\u52a8\u7a7a\u95f4\u5bf9\u9f50\u80fd\u4f7f\u667a\u80fd\u4f53\u5b8c\u6210\u66f4\u591a click\u3001type \u7b49\u5f15\u5bfc\u73af\u5883\u53d8\u5316\u7684\u52a8\u4f5c\uff0c\u89c2\u6d4b\u7a7a\u95f4\u5bf9\u9f50\u5219\u51cf\u5c11\u5927\u6a21\u578b\u8c03\u7528\u7684\u5b57\u7b26\u6570\u4e0e\u667a\u80fd\u4f53\u5b8c\u6210\u4efb\u52a1\u6240\u9700\u7684\u6b65\u6570\u3002<\/section>\n<section><a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-be5ec7a8d936502850a6e82ba5bfbd2e.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-be5ec7a8d936502850a6e82ba5bfbd2e.png\"><\/a><\/section>\n<section><strong>LLM-as-a-Judge\u00a0<\/strong><\/section>\n<section><\/section>\n<section>\u7814\u7a76\u56e2\u961f\u53d1\u73b0\uff0c\u667a\u80fd\u4f53\u7684\u51b3\u7b56\u884c\u4e3a\u6ce2\u52a8\u6027\u5f88\u5f3a\u3002\u7b80\u800c\u8a00\u4e4b\uff0c\u9762\u5bf9\u4e00\u4e2a\u76ee\u6807\uff0c\u667a\u80fd\u4f53\u6709\u4e00\u5b9a\u6982\u7387\u505a\u51fa\u6b63\u786e\u7684\u884c\u4e3a\u51b3\u65ad\uff0c\u4f46\u7531\u4e8e token \u9884\u6d4b\u7684\u968f\u673a\u6027\uff0c\u5b83\u53ef\u80fd\u505a\u51fa\u4e00\u4e9b\u9ad8\u6210\u672c\u3001\u4f4e\u56de\u62a5\u7684\u51b3\u5b9a\u3002\u8fd9\u4e5f\u5bfc\u81f4\u5b83\u5728\u540e\u7eed\u6b65\u9aa4\u4e2d\u96be\u4ee5\u7ea0\u6b63\u4e4b\u524d\u7684\u9519\u8bef\u800c\u5931\u8d25\u3002<\/section>\n<section><\/section>\n<section>\u4f8b\u5982\uff0c\u8981\u6c42\u667a\u80fd\u4f53\u5728\u67d0\u4e2a\u6700\u76f8\u5173\u7684\u8bdd\u9898\u4e0b\u53d1\u5e03\u5e16\u5b50\uff0c\u5355\u6b21 LLM \u8c03\u7528\u7684 AgentOccam \u5f80\u5f80\u8f7b\u7387\u5730\u9009\u62e9\u8bdd\u9898\uff0c\u672a\u8003\u8651\u300c\u6700\u76f8\u5173\u300d\u7684\u8981\u6c42\u3002<\/section>\n<section><\/section>\n<section>\u4e3a\u4e86\u89e3\u51b3\u6b64\u7c7b\u95ee\u9898\uff0c\u4ed6\u4eec\u5f15\u5bfc AgentOccam \u751f\u6210\u5355\u6b65\u5185\u6240\u6709\u53ef\u80fd\u7684\u884c\u52a8\uff0c\u8fd9\u7cfb\u5217\u884c\u52a8\u5c06\u4ea4\u4ed8\u53e6\u4e00\u4e2a Judge \u667a\u80fd\u4f53\uff08\u540c\u6837\u8c03\u7528 GPT-4-turbo\uff09\u51b3\u65ad\uff0c\u505a\u51fa\u6700\u5927\u5316\u56de\u62a5\u7684\u9009\u62e9\u3002<\/section>\n<section><\/section>\n<section><strong>\u4e0e\u590d\u5408\u7b56\u7565\u7ed3\u5408\u4f7f\u7528<\/strong><\/section>\n<section><\/section>\n<section>\u590d\u5408\u7b56\u7565\u4e2d\uff0c\u4e0e\u4efb\u52a1\u76f8\u5173\u7684\u7ecf\u9a8c\u53ef\u4ee5\u63d0\u5347\u667a\u80fd\u4f53\u6027\u80fd\u3002\u540c\u65f6\uff0c\u4e0d\u56e0\u4e3a\u52a0\u5165\u4e86\u66f4\u591a\u80cc\u666f\u77e5\u8bc6\u6270\u4e71\u51b3\u7b56\uff0c\u4e0d\u4f1a\u5f71\u54cd\u6cdb\u5316\u6027\uff0c\u80fd\u591f\u7ea0\u6b63\u9519\u8bef\u884c\u4e3a\u6a21\u5f0f\u3002<\/section>\n<section><\/section>\n<section>\u7531\u4e8e\u884c\u4e3a \/ \u89c2\u6d4b\u7a7a\u95f4\u5bf9\u9f50\u548c\u590d\u5408\u7b56\u7565\u65b9\u6cd5\u6b63\u4ea4\uff0c\u56e0\u6b64\u80fd\u7ed3\u5408\u5229\u7528\u3002\u8be5\u56e2\u961f\u8bd5\u9a8c\u5c06 AgentOccam \u4e0e 1\uff09SteP \u548c 2\uff09\u4e0a\u8ff0\u7684 LLM-as-a-Judge \u65b9\u6cd5\u8054\u5408\u4f7f\u7528\u3002<\/section>\n<section><\/section>\n<section>\u5bf9\u4e8e\u548c\u524d SOTA \u65b9\u6cd5 SteP \u8054\u5408\uff0c\u7531\u4e8e\u5b83\u5f15\u5165\u4eba\u7c7b\u7f16\u5199\u7684 WebArena \u4efb\u52a1\u653b\u7565\uff0c\u5728\u7ecf\u9a8c\u5bc6\u96c6\u578b\u4efb\u52a1\uff0c\u5982\u8d2d\u7269\u7f51\u9875\u4efb\u52a1\u4e2d\uff0c\u4eba\u7c7b\u64b0\u5199\u7684\u5f15\u5bfc\u6027\u7ecf\u9a8c\u5927\u5e45\u63d0\u5347\u4efb\u52a1\u6210\u529f\u7387\u3002<\/section>\n<section><\/section>\n<section>\u800c\u5728\u5e38\u8bc6\u6cdb\u5316\u5bc6\u96c6\u578b\u4efb\u52a1\uff0c\u5982\u793e\u4ea4\u7f51\u9875\u53d1\u5e16\u4efb\u52a1\u4e2d\uff0c\u4e0d\u76f8\u5173\u77e5\u8bc6\u51fa\u73b0\u4f1a\u9519\u8bef\u6270\u4e71\u667a\u80fd\u4f53\u51b3\u7b56\u3002\u5bf9\u4e8e LLM-as-a-Judge \u65b9\u6cd5\uff0cJudge \u89d2\u8272\u7684\u5f15\u5165\u4e0d\u5f71\u54cd\u667a\u80fd\u4f53\u7684\u6cdb\u5316\u6027\uff0c\u540c\u65f6\u7ea0\u6b63\u4e86\u667a\u80fd\u4f53\u4ed3\u4fc3\u51b3\u7b56\u7684\u9519\u8bef\u884c\u4e3a\u6a21\u5f0f\uff0c\u5728 WebArena \u4e0a\u8fdb\u4e00\u6b65\u63d0\u5347 2.6 \u7684\u7edd\u5bf9\u5206\u6570\u3002<\/section>\n<p>\u6587\u7ae0\u6765\u6e90\u4e8e\u4e92\u8054\u7f51:<a href=\"https:\/\/www.jiqizhixin.com\/articles\/2024-11-06-4\" target=\"_blank\">\u4e0d\u9760\u66f4\u590d\u6742\u7684\u7b56\u7565\uff0c\u4ec5\u51ed\u548c\u5927\u6a21\u578b\u8bad\u7ec3\u5bf9\u9f50\uff0c\u96f6\u6837\u672c\u96f6\u7ecf\u9a8c\u5355LLM\u8c03\u7528\uff0c\u6210\u4e3a\u7f51\u7edc\u4efb\u52a1\u667a\u80fd\u4f53\u65b0SOTA<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>\u6587\u7ae0\u6765\u6e90\u4e8e\u4e92\u8054\u7f51:\u4e0d\u9760\u66f4\u590d\u6742\u7684\u7b56\u7565\uff0c\u4ec5\u51ed [&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":[70,68],"class_list":["post-36537","post","type-post","status-publish","format-standard","hentry","category-news","tag-agent","tag-68"],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v26.4 - https:\/\/yoast.com\/wordpress\/plugins\/seo\/ -->\n<title>\u4e0d\u9760\u66f4\u590d\u6742\u7684\u7b56\u7565\uff0c\u4ec5\u51ed\u548c\u5927\u6a21\u578b\u8bad\u7ec3\u5bf9\u9f50\uff0c\u96f6\u6837\u672c\u96f6\u7ecf\u9a8c\u5355LLM\u8c03\u7528\uff0c\u6210\u4e3a\u7f51\u7edc\u4efb\u52a1\u667a\u80fd\u4f53\u65b0SOTA - \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=36537\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\/\/17aitech.com\/?p=36537\",\"url\":\"https:\/\/17aitech.com\/?p=36537\",\"name\":\"\u4e0d\u9760\u66f4\u590d\u6742\u7684\u7b56\u7565\uff0c\u4ec5\u51ed\u548c\u5927\u6a21\u578b\u8bad\u7ec3\u5bf9\u9f50\uff0c\u96f6\u6837\u672c\u96f6\u7ecf\u9a8c\u5355LLM\u8c03\u7528\uff0c\u6210\u4e3a\u7f51\u7edc\u4efb\u52a1\u667a\u80fd\u4f53\u65b0SOTA - \u4e00\u8d77AI\u6280\u672f\",\"isPartOf\":{\"@id\":\"https:\/\/17aitech.com\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\/\/17aitech.com\/?p=36537#primaryimage\"},\"image\":{\"@id\":\"https:\/\/17aitech.com\/?p=36537#primaryimage\"},\"thumbnailUrl\":\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-6057e6a045f6e7c4330c97faf2b5b94c.png\",\"datePublished\":\"2025-01-13T04:00:19+00:00\",\"author\":{\"@id\":\"https:\/\/17aitech.com\/#\/schema\/person\/60225458499e817ae0af73e67e440b9d\"},\"breadcrumb\":{\"@id\":\"https:\/\/17aitech.com\/?p=36537#breadcrumb\"},\"inLanguage\":\"zh-Hans\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/17aitech.com\/?p=36537\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"zh-Hans\",\"@id\":\"https:\/\/17aitech.com\/?p=36537#primaryimage\",\"url\":\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-6057e6a045f6e7c4330c97faf2b5b94c.png\",\"contentUrl\":\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-6057e6a045f6e7c4330c97faf2b5b94c.png\"},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/17aitech.com\/?p=36537#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"\u9996\u9875\",\"item\":\"https:\/\/17aitech.com\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"\u4e0d\u9760\u66f4\u590d\u6742\u7684\u7b56\u7565\uff0c\u4ec5\u51ed\u548c\u5927\u6a21\u578b\u8bad\u7ec3\u5bf9\u9f50\uff0c\u96f6\u6837\u672c\u96f6\u7ecf\u9a8c\u5355LLM\u8c03\u7528\uff0c\u6210\u4e3a\u7f51\u7edc\u4efb\u52a1\u667a\u80fd\u4f53\u65b0SOTA\"}]},{\"@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":"\u4e0d\u9760\u66f4\u590d\u6742\u7684\u7b56\u7565\uff0c\u4ec5\u51ed\u548c\u5927\u6a21\u578b\u8bad\u7ec3\u5bf9\u9f50\uff0c\u96f6\u6837\u672c\u96f6\u7ecf\u9a8c\u5355LLM\u8c03\u7528\uff0c\u6210\u4e3a\u7f51\u7edc\u4efb\u52a1\u667a\u80fd\u4f53\u65b0SOTA - \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=36537","schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/17aitech.com\/?p=36537","url":"https:\/\/17aitech.com\/?p=36537","name":"\u4e0d\u9760\u66f4\u590d\u6742\u7684\u7b56\u7565\uff0c\u4ec5\u51ed\u548c\u5927\u6a21\u578b\u8bad\u7ec3\u5bf9\u9f50\uff0c\u96f6\u6837\u672c\u96f6\u7ecf\u9a8c\u5355LLM\u8c03\u7528\uff0c\u6210\u4e3a\u7f51\u7edc\u4efb\u52a1\u667a\u80fd\u4f53\u65b0SOTA - \u4e00\u8d77AI\u6280\u672f","isPartOf":{"@id":"https:\/\/17aitech.com\/#website"},"primaryImageOfPage":{"@id":"https:\/\/17aitech.com\/?p=36537#primaryimage"},"image":{"@id":"https:\/\/17aitech.com\/?p=36537#primaryimage"},"thumbnailUrl":"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-6057e6a045f6e7c4330c97faf2b5b94c.png","datePublished":"2025-01-13T04:00:19+00:00","author":{"@id":"https:\/\/17aitech.com\/#\/schema\/person\/60225458499e817ae0af73e67e440b9d"},"breadcrumb":{"@id":"https:\/\/17aitech.com\/?p=36537#breadcrumb"},"inLanguage":"zh-Hans","potentialAction":[{"@type":"ReadAction","target":["https:\/\/17aitech.com\/?p=36537"]}]},{"@type":"ImageObject","inLanguage":"zh-Hans","@id":"https:\/\/17aitech.com\/?p=36537#primaryimage","url":"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-6057e6a045f6e7c4330c97faf2b5b94c.png","contentUrl":"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/11\/frc-6057e6a045f6e7c4330c97faf2b5b94c.png"},{"@type":"BreadcrumbList","@id":"https:\/\/17aitech.com\/?p=36537#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"\u9996\u9875","item":"https:\/\/17aitech.com\/"},{"@type":"ListItem","position":2,"name":"\u4e0d\u9760\u66f4\u590d\u6742\u7684\u7b56\u7565\uff0c\u4ec5\u51ed\u548c\u5927\u6a21\u578b\u8bad\u7ec3\u5bf9\u9f50\uff0c\u96f6\u6837\u672c\u96f6\u7ecf\u9a8c\u5355LLM\u8c03\u7528\uff0c\u6210\u4e3a\u7f51\u7edc\u4efb\u52a1\u667a\u80fd\u4f53\u65b0SOTA"}]},{"@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\/36537","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=36537"}],"version-history":[{"count":0,"href":"https:\/\/17aitech.com\/index.php?rest_route=\/wp\/v2\/posts\/36537\/revisions"}],"wp:attachment":[{"href":"https:\/\/17aitech.com\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=36537"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/17aitech.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=36537"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/17aitech.com\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=36537"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}