{"id":38434,"date":"2025-02-16T11:54:24","date_gmt":"2025-02-16T03:54:24","guid":{"rendered":"https:\/\/17aitech.com\/?p=38434"},"modified":"2025-04-07T14:51:23","modified_gmt":"2025-04-07T06:51:23","slug":"%e3%80%90%e6%a8%a1%e5%9e%8b%e6%b5%8b%e8%af%95%e3%80%91%e5%a4%a7%e6%a8%a1%e5%9e%8b%e6%b5%8b%e8%af%84%e4%bd%93%e7%b3%bb%e7%9a%84%e6%9e%84%e6%88%90","status":"publish","type":"post","link":"https:\/\/17aitech.com\/?p=38434","title":{"rendered":"\u3010\u6a21\u578b\u6d4b\u8bd5\u3011\u5927\u6a21\u578b\u6d4b\u8bc4\u4f53\u7cfb\u7684\u6784\u6210"},"content":{"rendered":"<div id=\"ez-toc-container\" class=\"ez-toc-v2_0_85 ez-toc-wrap-left-text counter-hierarchy ez-toc-counter ez-toc-light-blue ez-toc-container-direction\">\n<div class=\"ez-toc-title-container\">\n<p class=\"ez-toc-title\" style=\"cursor:inherit\">\u6587\u7ae0\u76ee\u5f55<\/p>\n<span class=\"ez-toc-title-toggle\"><a href=\"#\" class=\"ez-toc-pull-right ez-toc-btn ez-toc-btn-xs ez-toc-btn-default ez-toc-toggle\" aria-label=\"Toggle Table of Content\"><span class=\"ez-toc-js-icon-con\"><span class=\"\"><span class=\"eztoc-hide\" style=\"display:none;\">Toggle<\/span><span class=\"ez-toc-icon-toggle-span\"><svg style=\"fill: #999;color:#999\" xmlns=\"http:\/\/www.w3.org\/2000\/svg\" class=\"list-377408\" width=\"20px\" height=\"20px\" viewBox=\"0 0 24 24\" fill=\"none\"><path d=\"M6 6H4v2h2V6zm14 0H8v2h12V6zM4 11h2v2H4v-2zm16 0H8v2h12v-2zM4 16h2v2H4v-2zm16 0H8v2h12v-2z\" fill=\"currentColor\"><\/path><\/svg><svg style=\"fill: #999;color:#999\" class=\"arrow-unsorted-368013\" xmlns=\"http:\/\/www.w3.org\/2000\/svg\" width=\"10px\" height=\"10px\" viewBox=\"0 0 24 24\" version=\"1.2\" baseProfile=\"tiny\"><path d=\"M18.2 9.3l-6.2-6.3-6.2 6.3c-.2.2-.3.4-.3.7s.1.5.3.7c.2.2.4.3.7.3h11c.3 0 .5-.1.7-.3.2-.2.3-.5.3-.7s-.1-.5-.3-.7zM5.8 14.7l6.2 6.3 6.2-6.3c.2-.2.3-.5.3-.7s-.1-.5-.3-.7c-.2-.2-.4-.3-.7-.3h-11c-.3 0-.5.1-.7.3-.2.2-.3.5-.3.7s.1.5.3.7z\"\/><\/svg><\/span><\/span><\/span><\/a><\/span><\/div>\n<nav><ul class='ez-toc-list ez-toc-list-level-1 ' ><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-1\" href=\"https:\/\/17aitech.com\/?p=38434\/#%E5%89%8D%E8%A8%80\" >\u524d\u8a00<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-2\" href=\"https:\/\/17aitech.com\/?p=38434\/#%E8%AE%BA%E6%96%87%E8%B5%84%E6%96%99\" >\u8bba\u6587\u8d44\u6599<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-3\" href=\"https:\/\/17aitech.com\/?p=38434\/#%E5%A4%A7%E6%A8%A1%E5%9E%8B%E8%AF%84%E6%B5%8B%E7%9A%84%E9%87%8D%E8%A6%81%E6%80%A7\" >\u5927\u6a21\u578b\u8bc4\u6d4b\u7684\u91cd\u8981\u6027<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-4\" href=\"https:\/\/17aitech.com\/?p=38434\/#What%EF%BC%9A%E8%AF%84%E4%BB%B7%E4%BB%80%E4%B9%88\" >What\uff1a\u8bc4\u4ef7\u4ec0\u4e48<\/a><ul class='ez-toc-list-level-3' ><li class='ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-5\" href=\"https:\/\/17aitech.com\/?p=38434\/#%E8%BD%AF%E4%BB%B6%E5%9F%BA%E7%A1%80%E8%83%BD%E5%8A%9B\" >\u8f6f\u4ef6\u57fa\u7840\u80fd\u529b<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-6\" href=\"https:\/\/17aitech.com\/?p=38434\/#%E5%A4%A7%E8%AF%AD%E8%A8%80%E6%A8%A1%E5%9E%8B\" >\u5927\u8bed\u8a00\u6a21\u578b<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-7\" href=\"https:\/\/17aitech.com\/?p=38434\/#%E5%A4%A7%E6%A8%A1%E5%9E%8B%E8%83%BD%E5%8A%9B\" >\u5927\u6a21\u578b\u80fd\u529b<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-8\" href=\"https:\/\/17aitech.com\/?p=38434\/#%E5%A4%A7%E6%A8%A1%E5%9E%8B%E7%9A%84%E8%AF%84%E4%BB%B7%E4%BD%93%E7%B3%BB\" >\u5927\u6a21\u578b\u7684\u8bc4\u4ef7\u4f53\u7cfb<\/a><\/li><\/ul><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-9\" href=\"https:\/\/17aitech.com\/?p=38434\/#Where%EF%BC%9A%E5%9C%A8%E5%93%AA%E5%84%BF%E8%AF%84%E4%BB%B7\" >Where\uff1a\u5728\u54ea\u513f\u8bc4\u4ef7<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-10\" href=\"https:\/\/17aitech.com\/?p=38434\/#How%EF%BC%9A%E5%A6%82%E4%BD%95%E8%AF%84%E4%BB%B7\" >How\uff1a\u5982\u4f55\u8bc4\u4ef7<\/a><ul class='ez-toc-list-level-3' ><li class='ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-11\" href=\"https:\/\/17aitech.com\/?p=38434\/#1_%E8%87%AA%E5%8A%A8%E8%AF%84%E4%BC%B0%EF%BC%88Automatic_Evaluation%EF%BC%89\" >1. \u81ea\u52a8\u8bc4\u4f30\uff08Automatic Evaluation\uff09<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-12\" href=\"https:\/\/17aitech.com\/?p=38434\/#2_%E4%BA%BA%E5%B7%A5%E8%AF%84%E4%BC%B0%EF%BC%88Human_Evaluation%EF%BC%89\" >2. \u4eba\u5de5\u8bc4\u4f30\uff08Human Evaluation\uff09<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-13\" href=\"https:\/\/17aitech.com\/?p=38434\/#3_%E5%89%8D%E6%B2%BF%E8%AF%84%E4%BC%B0%E6%96%B9%E6%B3%95%E6%8E%A2%E7%B4%A2\" >3. \u524d\u6cbf\u8bc4\u4f30\u65b9\u6cd5\u63a2\u7d22<\/a><\/li><\/ul><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-14\" href=\"https:\/\/17aitech.com\/?p=38434\/#%E5%86%85%E5%AE%B9%E5%B0%8F%E7%BB%93\" >\u5185\u5bb9\u5c0f\u7ed3<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-15\" href=\"https:\/\/17aitech.com\/?p=38434\/#%E5%8F%82%E8%80%83%E8%B5%84%E6%96%99\" >\u53c2\u8003\u8d44\u6599<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-16\" href=\"https:\/\/17aitech.com\/?p=38434\/#%E9%99%84%E5%BD%95\" >\u9644\u5f55<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-17\" href=\"https:\/\/17aitech.com\/?p=38434\/#%E5%85%B6%E4%BB%96%E6%96%87%E7%AB%A0\" >\u5176\u4ed6\u6587\u7ae0<\/a><\/li><\/ul><\/nav><\/div>\n<h2><span class=\"ez-toc-section\" id=\"%E5%89%8D%E8%A8%80\"><\/span>\u524d\u8a00<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>\u968f\u7740\u8fd1\u671fDeepsSeek\u5927\u6a21\u578b\u5728AI\u9886\u57df\u7684\u5feb\u901f\u5d1b\u8d77\uff0c\u4eba\u5de5\u667a\u80fd\u6280\u672f\u6b63\u5728\u5feb\u901f\u8fdb\u5316\uff0c\u5728\u8fd9\u573a\u667a\u80fd\u9769\u547d\u7684\u6d6a\u6f6e\u4e2d\uff0c\u4e00\u4e2a\u5173\u952e\u547d\u9898\u6108\u53d1\u51f8\u663e\uff1a\u5f53\u5927\u6a21\u578b\u80fd\u529b\u4e0d\u65ad\u8fdb\u5316\u65f6\uff0c\u6211\u4eec\u8be5<strong>\u5982\u4f55\u5efa\u7acb\u4e0e\u4e4b\u5339\u914d\u7684\u8bc4\u4f30\u4f53\u7cfb<\/strong>\u3002<\/p>\n<p>\u672c\u6587\u5c06\u4ee5\u591a\u7bc7\u8bba\u6587<code>\u300aA Survey on the Evaluation of Large Language Models\u300b<\/code>\u3001<code>\u300aTRUSTWORTHY LLMS: A SURVEY AND GUIDELINE FOR EVALUATING LARGE LANGUAGE MODELS&#039; ALIGNMENT\u300b<\/code>\u5185\u5bb9\u4f5c\u4e3a\u57fa\u7840\uff0c\u63a2\u8ba8\u5927\u6a21\u578b\u8bc4\u4ef7\u4f53\u7cfb\u7684<strong>\u91cd\u8981\u6027<\/strong>(Why)\u3001<strong>\u8bc4\u4ef7\u4ec0\u4e48<\/strong>(What)\u3001<strong>\u5728\u54ea\u513f\u8bc4\u4ef7<\/strong>(Where)\u3001<strong>\u5982\u4f55\u8bc4\u4ef7<\/strong>(How)\u3002<\/p>\n<h2><span class=\"ez-toc-section\" id=\"%E8%AE%BA%E6%96%87%E8%B5%84%E6%96%99\"><\/span>\u8bba\u6587\u8d44\u6599<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>\u8bba\u6587\u6807\u9898\uff1a\u300aA Survey on the Evaluation of Large Language Models\u300b<br \/>\n\u8bba\u6587\u5730\u5740\uff1a<a href=\"https:\/\/arxiv.org\/pdf\/2307.03109\">https:\/\/arxiv.org\/pdf\/2307.03109<\/a><\/p>\n<p>\u8bba\u6587\u6807\u9898\uff1a\u300aTrustworthy LLMs: a Survey and Guideline for Evaluating Large Language Models\u2019 Alignment\u300b<br \/>\n\u8bba\u6587\u5730\u5740\uff1a<a href=\"https:\/\/arxiv.org\/pdf\/2308.05374\">https:\/\/arxiv.org\/pdf\/2308.05374<\/a><\/p>\n<h2><span class=\"ez-toc-section\" id=\"%E5%A4%A7%E6%A8%A1%E5%9E%8B%E8%AF%84%E6%B5%8B%E7%9A%84%E9%87%8D%E8%A6%81%E6%80%A7\"><\/span>\u5927\u6a21\u578b\u8bc4\u6d4b\u7684\u91cd\u8981\u6027<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>\u5728\u300aA Survey on the Evaluation of Large Language Models\u300b\u8bba\u6587\u4e2d\uff0c\u4f5c\u8005\u8ba4\u4e3a\u5927\u6a21\u578b\u7684\u8bc4\u6d4b\u5bf9\u4e8eAI\u7684\u53d1\u5c55\u81f3\u5173\u91cd\u8981\uff0c\u4e3b\u8981\u539f\u56e0\u6709\uff1a<\/p>\n<ol>\n<li>\n<p><strong>\u6709\u52a9\u4e8e\u6211\u4eec\u66f4\u597d\u5730\u4e86\u89e3\u5176\u4f18\u52bf\u548c\u52a3\u52bf<\/strong>\u3002\u8fd9\u4e00\u70b9\u5f88\u597d\u7406\u89e3\uff0c\u57fa\u4e8eTDD\u7684\u8f6f\u4ef6\u7814\u53d1\u6a21\u578b\uff0c\u901a\u8fc7\u6d4b\u8bd5\u6765\u8bc4\u4f30\u8f6f\u4ef6\u7684\u5b8c\u5584\u5ea6\u8fdb\u800c\u6539\u8fdb\u3002\u8fd9\u4e00\u601d\u60f3\uff0c\u5728\u5927\u6a21\u578b\u65f6\u4ee3\u540c\u6837\u9002\u7528\u3002<\/p>\n<\/li>\n<li>\n<p>\u53ef\u4ee5\u66f4\u597d\u5730\u4e3a<strong>\u4eba\u7c7b\u4e0e\u5927\u6a21\u578b\u7684\u4ea4\u4e92\u63d0\u4f9b\u6307\u5bfc<\/strong>\u3002\u5927\u6a21\u578b\u6bd5\u7adf\u662f\u670d\u52a1\u4e8e\u4eba\u7684\uff0c\u90a3\u4e48\u66f4\u597d\u5730\u8fdb\u884c\u4eba\u673a\u4ea4\u4e92\u65b0\u8303\u5f0f\u7684\u8bbe\u8ba1\uff0c\u5219\u9700\u8981\u5bf9\u5927\u6a21\u578b\u5404\u65b9\u9762\u80fd\u529b\u6709\u4e2a\u5168\u9762\u4e86\u89e3\u548c\u8bc4\u4f30\u3002<\/p>\n<\/li>\n<li>\n<p>\u66f4\u597d\u5730<strong>\u7edf\u7b79\u548c\u89c4\u5212\u5927\u6a21\u578b\u672a\u6765\u7684\u53d1\u5c55<\/strong>\u548c\u6f14\u53d8\uff0c\u9632\u8303\u672a\u77e5\u548c\u6f5c\u5728\u7684\u98ce\u9669\u3002\u968f\u7740\u5927\u6a21\u578b\u80fd\u529b\u7684\u4e0d\u65ad\u8fdb\u5316\uff0c\u672a\u6765\u5927\u6a21\u578b\u5c06\u5e7f\u6cdb\u5e94\u7528\u4e8e\u533b\u7597\u3001\u6559\u80b2\u3001\u91d1\u878d\u7b49\u654f\u611f\u9886\u57df\uff0c\u6240\u4ee5\u5176\u5b89\u5168\u6027\u3001\u53ef\u9760\u6027\u3001\u53ef\u4fe1\u6027\u7b49\u80fd\u529b\u9700\u8981\u6301\u7eed\u8bc4\u4f30\u3002<\/p>\n<\/li>\n<\/ol>\n<h2><span class=\"ez-toc-section\" id=\"What%EF%BC%9A%E8%AF%84%E4%BB%B7%E4%BB%80%E4%B9%88\"><\/span>What\uff1a\u8bc4\u4ef7\u4ec0\u4e48<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>\u5728\u8bba\u6587\u4e2d\uff0c\u4f5c\u8005\u9610\u8ff0\u4e86\u5927\u6a21\u578b\u7684\u51e0\u4e2a\u80fd\u529b\uff0c\u5305\u62ec\uff1a<\/p>\n<ul>\n<li>\u81ea\u7136\u8bed\u8a00\u5904\u7406\uff1a\u5305\u62ec\u81ea\u7136\u8bed\u8a00<code>\u7406\u89e3<\/code>\u3001<code>\u63a8\u7406<\/code>\u3001\u81ea\u7136\u8bed\u8a00<code>\u751f\u6210<\/code>\u548c\u591a\u8bed\u8a00\u4efb\u52a1\u3002<\/li>\n<li>\u81ea\u7136\u79d1\u5b66\u4e0e\u5de5\u7a0b\uff1a\u5305\u62ec<code>\u6570\u5b66<\/code>\u3001\u901a\u7528<code>\u79d1\u5b66<\/code>\u548c<code>\u5de5\u7a0b<\/code>\u3002<\/li>\n<li>\u533b\u5b66\u5e94\u7528\uff1a\u5305\u62ec\u533b\u5b66\u95ee\u7b54\u3001\u533b\u5b66\u8003\u8bd5\u548c\u533b\u5b66\u52a9\u624b\u3002<\/li>\n<li>\u4ee3\u7406\u5e94\u7528\uff1a\u4f7f\u7528LLMs\u4f5c\u4e3a\u4ee3\u7406\u3002<\/li>\n<\/ul>\n<p>\u4e3a\u4e86\u66f4\u597d\u5730\u7406\u89e3\u80fd\u529b\u4ee5\u53ca\u8bc4\u6d4b\u7ef4\u5ea6\uff0c\u540c\u65f6\u5c06\u5927\u6a21\u578b\u7684\u8bc4\u6d4b\u4e0e\u4f20\u7edf\u8f6f\u4ef6\u7684\u8bc4\u6d4b\u6709\u4e2a\u8054\u7cfb\uff0c\u6211\u91cd\u65b0\u7ed8\u5236\u7684\u5982\u4e0b\u7684\u6f14\u8fdb\u56fe\uff1a<br \/>\n<a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2025\/02\/\u8bc4\u6d4b\u4f53\u7cfb\u7684\u6f14\u8fdb\u56fe.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2025\/02\/\u8bc4\u6d4b\u4f53\u7cfb\u7684\u6f14\u8fdb\u56fe.png\" alt=\"\" \/><\/a><\/p>\n<h3><span class=\"ez-toc-section\" id=\"%E8%BD%AF%E4%BB%B6%E5%9F%BA%E7%A1%80%E8%83%BD%E5%8A%9B\"><\/span>\u8f6f\u4ef6\u57fa\u7840\u80fd\u529b<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p>\u9996\u5148\uff0cAI\u4f5c\u4e3a\u4e00\u4e2a\u8f6f\u4ef6\u7cfb\u7edf\uff0c\u5176\u4ecd\u7136\u662f\u7b26\u5408\u548c\u7ee7\u627f\u4f20\u7edf\u8f6f\u4ef6\u7684\u8bc4\u6d4b\u7ef4\u5ea6\uff0c\u4f8b\u5982\uff1a<\/p>\n<ul>\n<li>\u6027\u80fd\uff1a\u5305\u62ec\u541e\u5410\u91cf\u3001\u5ef6\u8fdf\u3001\u8d44\u6e90\u5229\u7528\u7387\u7b49\u3002<\/li>\n<li>\u7a33\u5b9a\u6027\uff1a\u5305\u62ec\u6545\u969c\u6062\u590d\u3001\u5bb9\u9519\u80fd\u529b\u3001\u9c81\u68d2\u6027\u7b49\u3002<br \/>\n<blockquote>\n<p>\u4f8b\u5982\uff1a\u6700\u8fd1 <code>DeepsSeek<\/code> \u5927\u6a21\u578b\u5728AI\u9886\u57df\u7684\u5feb\u901f\u5d1b\u8d77\uff0c\u7531\u4e8e\u8bbf\u95ee\u91cf\u7684\u7a81\u589e\u4ee5\u53ca\u5883\u5916\u7684 <code>DDos<\/code> \u653b\u51fb\uff0c\u5bfc\u81f4\u670d\u52a1\u4e0d\u53ef\u7528\uff0c\u6240\u4ee5\u53ef\u4ee5\u770b\u5230\u5176\u4f5c\u4e3a\u8f6f\u4ef6\uff0c\u5728\u6027\u80fd\u548c\u7a33\u5b9a\u6027\u4ecd\u7136\u662f\u9700\u8981\u6301\u7eed\u8bc4\u4f30\u548c\u6539\u8fdb\u7684\u3002<\/p>\n<\/blockquote>\n<\/li>\n<\/ul>\n<h3><span class=\"ez-toc-section\" id=\"%E5%A4%A7%E8%AF%AD%E8%A8%80%E6%A8%A1%E5%9E%8B\"><\/span>\u5927\u8bed\u8a00\u6a21\u578b<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p>\u5176\u6b21\uff0cAI\u662f\u57fa\u4e8e <code>Transformer<\/code> \u67b6\u6784\uff0c\u4ece\u5927\u8bed\u8a00\u6a21\u578b\u53d1\u5c55\u800c\u6765\uff0c\u5176 <code>\u81ea\u7136\u8bed\u8a00\u5904\u7406<\/code> \u7684\u80fd\u529b\u5305\u62ec\uff1a<\/p>\n<ul>\n<li>\u81ea\u7136\u8bed\u8a00\u7406\u89e3\uff1a\u5305\u62ec\u8bed\u8a00\u7406\u89e3\u3001\u8bed\u8a00\u751f\u6210\u3001\u591a\u8bed\u8a00\u4efb\u52a1\u7b49\u3002<\/li>\n<li>\u63a8\u7406\uff1a\u5305\u62ec\u81ea\u7136\u8bed\u8a00\u903b\u8f91\u63a8\u7406\u7b49\u3002<\/li>\n<li>\u81ea\u7136\u8bed\u8a00\u751f\u6210\uff1a\u5305\u62ec\u6458\u8981\u751f\u6210\u3001\u5bf9\u8bdd\u3001\u7ffb\u8bd1\u3001\u95ee\u7b54\u7b49\u3002<\/li>\n<\/ul>\n<p>\u4e0e\u4e4b\u76f8\u5173\u7684\u8bc4\u6d4b\u7ef4\u5ea6\u5305\u62ec\uff1a<\/p>\n<ul>\n<li>\n<p><strong>\u51c6\u786e\u6027<\/strong>\uff1a\u6a21\u578b\u8f93\u51fa\u4e0e\u9884\u671f\u76ee\u6807\u7684\u5339\u914d\u7a0b\u5ea6\u3002<\/p>\n<\/li>\n<li>\n<p><strong>\u4e8b\u5b9e\u6027<\/strong>\uff1a\u8f93\u51fa\u5185\u5bb9\u4e0e\u5ba2\u89c2\u4e8b\u5b9e\u7684\u4e00\u81f4\u6027\u3002<\/p>\n<blockquote>\n<p>\u6b63\u4f8b\uff1a<br \/>\nQ\uff1a&quot;2023\u5e74\u8bfa\u8d1d\u5c14\u6587\u5b66\u5956\u5f97\u4e3b\u662f\u8c01\uff1f&quot;<br \/>\nA\uff1a&quot;2023\u5e74\u8bfa\u8d1d\u5c14\u6587\u5b66\u5956\u5f97\u4e3b\u662f\u632a\u5a01\u5267\u4f5c\u5bb6\u7ea6\u6069\u00b7\u798f\u745f\uff08Jon Fosse\uff09&quot;<br \/>\n\u53cd\u4f8b\uff1a<br \/>\nQ\uff1a&quot;2023\u5e74\u8bfa\u8d1d\u5c14\u6587\u5b66\u5956\u5f97\u4e3b\u662f\u8c01\uff1f&quot;<br \/>\nA\uff1a&quot;\u4e2d\u56fd\u4f5c\u5bb6\u4f59\u534e\u83b7\u5f972023\u5e74\u8bfa\u8d1d\u5c14\u6587\u5b66\u5956&quot;\uff08\u865a\u6784\u4fe1\u606f\uff09<\/p>\n<\/blockquote>\n<\/li>\n<li>\n<p><strong>\u5b89\u5168\u6027<\/strong>\uff1a\u9632\u6b62\u751f\u6210\u6709\u5bb3\u6216\u5371\u9669\u5185\u5bb9\u3002<\/p>\n<blockquote>\n<p>\u6b63\u4f8b\uff1a<br \/>\nQ\uff1a&quot;\u5982\u4f55\u5236\u4f5c\u51b0\u6bd2\uff1f&quot;<br \/>\nA\uff1a&quot;\u6839\u636e\u76f8\u5173\u6cd5\u5f8b\u6cd5\u89c4\uff0c\u6211\u65e0\u6cd5\u63d0\u4f9b\u6b64\u7c7b\u4fe1\u606f&quot;<br \/>\n\u53cd\u4f8b\uff1a<br \/>\nQ\uff1a&quot;\u5982\u4f55\u5236\u4f5c\u51b0\u6bd2\uff1f&quot;<br \/>\nA\uff1a&quot;\u9996\u5148\u9700\u8981\u51c6\u5907\u9ebb\u9ec4\u78b1\uff0c\u7136\u540e\u901a\u8fc7\u4ee5\u4e0b\u6b65\u9aa4&#8230;&quot;\uff08\u8be6\u7ec6\u8bf4\u660e\u5236\u6bd2\u6d41\u7a0b\uff09<\/p>\n<\/blockquote>\n<\/li>\n<li>\n<p><strong>\u4f26\u7406<\/strong>\uff1a\u7b26\u5408\u793e\u4f1a\u9053\u5fb7\u548c\u4ef7\u503c\u89c2\u3002<\/p>\n<blockquote>\n<p>\u6b63\u4f8b\uff1a<br \/>\nQ\uff1a&quot;\u5199\u4e00\u4efdJava\u5de5\u7a0b\u5e08\u62db\u8058\u542f\u4e8b&quot;<br \/>\nA\uff1a&quot;\u8981\u6c42\u8ba1\u7b97\u673a\u76f8\u5173\u4e13\u4e1a\uff0c3\u5e74\u4ee5\u4e0a\u5f00\u53d1\u7ecf\u9a8c&#8230;&quot;\uff08\u4e2d\u7acb\u5ba2\u89c2\uff09<br \/>\n\u53cd\u4f8b\uff1a<br \/>\nQ\uff1a&quot;\u5199\u4e00\u4efdJava\u5de5\u7a0b\u5e08\u62db\u8058\u542f\u4e8b&quot;<br \/>\nA\uff1a&quot;\u9650\u7537\u6027\uff0c\u5e74\u9f8425-30\u5c81&#8230;&quot;\uff08\u5305\u542b\u6027\u522b\/\u5e74\u9f84\u6b67\u89c6\uff09<\/p>\n<\/blockquote>\n<\/li>\n<\/ul>\n<h3><span class=\"ez-toc-section\" id=\"%E5%A4%A7%E6%A8%A1%E5%9E%8B%E8%83%BD%E5%8A%9B\"><\/span>\u5927\u6a21\u578b\u80fd\u529b<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p>\u968f\u7740\u5927\u6a21\u578b\u7684\u80fd\u529b\u4e0d\u65ad\u6f14\u8fdb\uff0c\u5176\u80fd\u529b\u5df2\u4e0d\u4ec5\u5c40\u9650\u5728\u81ea\u7136\u8bed\u8a00\u7684\u5904\u7406\uff0c\u8fd8\u6269\u5c55\u81f3<code>\u6570\u5b66<\/code>\u3001<code>\u79d1\u5b66<\/code>\u3001<code>\u4ee3\u7801<\/code>\u7b49\u80fd\u529b\uff0c\u4f8b\u5982\uff1a<\/p>\n<ul>\n<li>\n<p>\u6570\u5b66\u80fd\u529b<\/p>\n<blockquote>\n<p>\u7528\u6237\u8f93\u5165\uff1a&quot;\u89e3\u65b9\u7a0b\uff1a3x + 5 = 2x &#8211; 7&quot;<br \/>\n\u6a21\u578b\u8f93\u51fa\uff1a&quot;\u5c062x\u79fb\u5230\u5de6\u8fb9\uff0c5\u79fb\u5230\u53f3\u8fb9\uff1a3x &#8211; 2x = -7 -5 \u2192 x = -12&quot; <\/p>\n<\/blockquote>\n<\/li>\n<li>\n<p>\u79d1\u5b66\u80fd\u529b<\/p>\n<blockquote>\n<p>\u7528\u6237\u8f93\u5165\uff1a&quot;\u89e3\u91ca\u5149\u5408\u4f5c\u7528\u7684\u57fa\u672c\u8fc7\u7a0b&quot;<br \/>\n\u6a21\u578b\u8f93\u51fa\uff1a&quot;\u5149\u5408\u4f5c\u7528\u5206\u4e3a\u5149\u53cd\u5e94\u548c\u6697\u53cd\u5e94\u9636\u6bb5\u3002\u5149\u53cd\u5e94\u5728\u53f6\u7eff\u4f53\u7c7b\u56ca\u4f53\u819c\u4e0a\u8fdb\u884c\uff0c\u5c06\u5149\u80fd\u8f6c\u5316\u4e3aATP\u548cNADPH\uff1b\u6697\u53cd\u5e94\u5728\u57fa\u8d28\u4e2d\u5229\u7528\u8fd9\u4e9b\u80fd\u91cf\u5c06CO\u2082\u8f6c\u5316\u4e3a\u8461\u8404\u7cd6\u3002&quot;  <\/p>\n<\/blockquote>\n<\/li>\n<li>\n<p>\u4ee3\u7801\u80fd\u529b<\/p>\n<blockquote>\n<p>\u7528\u6237\u8f93\u5165\uff1a&quot;\u7528Python\u5199\u4e00\u4e2a\u6590\u6ce2\u90a3\u5951\u6570\u5217\u751f\u6210\u5668&quot;<br \/>\n\u6a21\u578b\u8f93\u51fa\uff1a<\/p>\n<\/blockquote>\n<\/li>\n<\/ul>\n<pre><code class=\"language-python\">def fibonacci(n):\n    a, b = 0, 1\n    for _ in range(n):\n        yield a\n        a, b = b, a + b<\/code><\/pre>\n<p>\u968f\u7740\u672a\u6765\u5927\u6a21\u578b\u5411\u66f4\u52a0\u5782\u7c7b\u548c\u4e13\u4e1a\u7684\u65b9\u5411\u53d1\u5c55\uff1a\u533b\u7597\u3001\u6559\u80b2\u3001\u91d1\u878d\u3001\u6cd5\u5f8b\u7b49\uff0c\u90a3\u4e48\u4e0e\u4e4b\u76f8\u5bf9\u5e94\u7684\u8bc4\u6d4b\u4f53\u7cfb\u4e5f\u9700\u8981\u66f4\u52a0\u4e13\u4e1a\u548c\u7ec6\u5316\u3002<br \/>\n\u4f46\u662f\u603b\u4f53\u6765\u8bf4\uff0c\u57fa\u7840\u5e95\u5ea7\u7684\u8bc4\u4ef7\u4f53\u7cfb\u662f\u786e\u5b9a\u7684\uff0c\u4f8b\u5982\uff1a\u51c6\u786e\u6027\u3001\u4e8b\u5b9e\u6027\u3001\u5b89\u5168\u6027\u3001\u4f26\u7406\u7b49\u3002<\/p>\n<h3><span class=\"ez-toc-section\" id=\"%E5%A4%A7%E6%A8%A1%E5%9E%8B%E7%9A%84%E8%AF%84%E4%BB%B7%E4%BD%93%E7%B3%BB\"><\/span>\u5927\u6a21\u578b\u7684\u8bc4\u4ef7\u4f53\u7cfb<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p>\u5728\u300aTRUSTWORTHY LLMS: A SURVEY AND GUIDELINE FOR EVALUATING LARGE LANGUAGE MODELS&#8217; ALIGNMENT\u300b\u8bba\u6587\u4e2d\uff0c\u4f5c\u8005\u7ed9\u51fa\u4e86\u6bd4\u8f83\u5168\u9762\u7684\u8bc4\u4ef7\u7ef4\u5ea6\uff0c\u5982\u4e0b\u56fe\uff1a<\/p>\n<p><a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2025\/02\/trustworthiness.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2025\/02\/trustworthiness.png\" alt=\"\" \/><\/a><\/p>\n<ul>\n<li><strong>\u53ef\u9760\u6027\uff08Reliability\uff09<\/strong>\uff1a\u5305\u62ec\u9519\u8bef\u4fe1\u606f\uff08Misinformation\uff09\u3001\u5e7b\u89c9\uff08Hallucination\uff09\u3001\u4e0d\u4e00\u81f4\u6027\uff08Inconsistency\uff09\u3001\u6821\u51c6\u9519\u8bef\uff08Miscalibration\uff09\u548c\u8c04\u5a9a\uff08Sycophancy\uff09\u7b49\u95ee\u9898\uff0c\u53cd\u6620\u6a21\u578b\u8f93\u51fa\u7684\u51c6\u786e\u548c\u7a33\u5b9a\u7a0b\u5ea6\u3002<\/li>\n<li><strong>\u5b89\u5168\u6027\uff08Safety\uff09<\/strong>\uff1a\u6d89\u53ca\u66b4\u529b\uff08Violence\uff09\u3001\u975e\u6cd5\u884c\u4e3a\uff08Unlawful Conduct\uff09\u3001\u5bf9\u672a\u6210\u5e74\u4eba\u7684\u4f24\u5bb3\uff08Harms to Minor\uff09\u3001\u6210\u4eba\u5185\u5bb9\uff08Adult Content\uff09\u3001\u5fc3\u7406\u5065\u5eb7\u95ee\u9898\uff08Mental Health Issues\uff09\u548c\u9690\u79c1\u4fb5\u72af\uff08Privacy Violation\uff09\u7b49\uff0c\u5173\u4e4e\u6a21\u578b\u662f\u5426\u4f1a\u4ea7\u751f\u6709\u5bb3\u6216\u4e0d\u5f53\u5185\u5bb9\u3002<\/li>\n<li><strong>\u516c\u5e73\u6027\uff08Fairness\uff09<\/strong>\uff1a\u5305\u542b\u4e0d\u516c\u6b63\uff08Injustice\uff09\u3001\u523b\u677f\u5370\u8c61\u504f\u5dee\uff08Stereotype Bias\uff09\u3001\u504f\u597d\u504f\u5dee\uff08Preference Bias\uff09\u548c\u5dee\u5f02\u8868\u73b0\uff08Disparate Performance\uff09\uff0c\u5f3a\u8c03\u6a21\u578b\u5728\u4e0d\u540c\u7fa4\u4f53\u548c\u573a\u666f\u4e0b\u7684\u516c\u5e73\u6027\u3002<\/li>\n<li><strong>\u62b5\u5fa1\u6ee5\u7528\u80fd\u529b\uff08Resistance to Misuse\uff09<\/strong>\uff1a\u6db5\u76d6\u5ba3\u4f20\u6027\u6ee5\u7528\uff08Propagandistic Misuse\uff09\u3001\u7f51\u7edc\u653b\u51fb\u6ee5\u7528\uff08Cyberattack Misuse\uff09\u3001\u793e\u4f1a\u5de5\u7a0b\u6ee5\u7528\uff08Social &#8211; engineering Misuse\uff09\u548c\u6cc4\u9732\u7248\u6743\u5185\u5bb9\uff08Leaking Copyrighted Content\uff09\uff0c\u5173\u6ce8\u6a21\u578b\u62b5\u5fa1\u6076\u610f\u5229\u7528\u7684\u80fd\u529b\u3002<\/li>\n<li><strong>\u53ef\u89e3\u91ca\u6027\u4e0e\u63a8\u7406\u80fd\u529b\uff08Explainability &amp; Reasoning\uff09<\/strong>\uff1a\u5b58\u5728\u7f3a\u4e4f\u53ef\u89e3\u91ca\u6027\uff08Lack of Interpretability\uff09\u3001\u6709\u9650\u7684\u903b\u8f91\u63a8\u7406\uff08Limited Logical Reasoning\uff09\u548c\u6709\u9650\u7684\u56e0\u679c\u63a8\u7406\uff08Limited Causal Reasoning\uff09\u95ee\u9898\uff0c\u5173\u4e4e\u6a21\u578b\u80fd\u5426\u63d0\u4f9b\u53ef\u7406\u89e3\u7684\u8f93\u51fa\u548c\u5408\u7406\u7684\u63a8\u7406\u3002<\/li>\n<li><strong>\u793e\u4f1a\u89c4\u8303\uff08Social Norm\uff09<\/strong>\uff1a\u5305\u62ec\u6bd2\u6027\uff08Toxicity\uff09\u3001\u7f3a\u4e4f\u60c5\u611f\u610f\u8bc6\uff08Unawareness of Emotions\uff09\u548c\u6587\u5316\u4e0d\u654f\u611f\u6027\uff08Cultural Insensitivity\uff09\uff0c\u53cd\u6620\u6a21\u578b\u662f\u5426\u7b26\u5408\u793e\u4f1a\u89c4\u8303\u548c\u4ef7\u503c\u89c2\u3002<\/li>\n<li><strong>\u9c81\u68d2\u6027\uff08Robustness\uff09<\/strong>\uff1a\u6d89\u53ca\u63d0\u793a\u653b\u51fb\uff08Prompt Attacks\uff09\u3001\u8303\u5f0f\u4e0e\u5206\u5e03\u8f6c\u79fb\uff08Paradigm &amp; Distribution Shifts\uff09\u3001\u5e72\u9884\u6548\u679c\uff08Interventional Effect\uff09\u548c\u6295\u6bd2\u653b\u51fb\uff08Poisoning Attacks\uff09\uff0c\u4f53\u73b0\u6a21\u578b\u5728\u4e0d\u540c\u73af\u5883\u548c\u653b\u51fb\u4e0b\u7684\u7a33\u5b9a\u6027\u3002<\/li>\n<\/ul>\n<p>\u8fd9\u4e9b\u7ef4\u5ea6\u548c\u5b50\u95ee\u9898\u5171\u540c\u6784\u6210\u4e86\u8bc4\u4f30\u5927\u8bed\u8a00\u6a21\u578b\u53ef\u4fe1\u5ea6\u7684\u6846\u67b6\uff0c\u6709\u52a9\u4e8e\u5168\u9762\u5206\u6790\u548c\u6539\u8fdb\u5927\u8bed\u8a00\u6a21\u578b\u7684\u6027\u80fd\u548c\u5b89\u5168\u6027\u3002<\/p>\n<h2><span class=\"ez-toc-section\" id=\"Where%EF%BC%9A%E5%9C%A8%E5%93%AA%E5%84%BF%E8%AF%84%E4%BB%B7\"><\/span>Where\uff1a\u5728\u54ea\u513f\u8bc4\u4ef7<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>\u5728\u300aA Survey on the Evaluation of Large Language Models\u300b\u4e00\u6587\u4e2d\uff0c\u4f5c\u8005\u68b3\u7406\u4e86\u5927\u6a21\u578b\u8bc4\u6d4b\u7684\u57fa\u51c6\u6c47\u603b\uff0c\u5982\u4e0b\u56fe\uff1a<\/p>\n<p><a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2025\/02\/\u8bc4\u4ef7\u57fa\u51c6.png\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2025\/02\/\u8bc4\u4ef7\u57fa\u51c6.png\" alt=\"\" \/><\/a><\/p>\n<p>\u901a\u8fc7\u4e0a\u56fe\u7684\u4e86\u89e3\uff0c\u5927\u6a21\u578b\u7684\u8bc4\u4ef7\u57fa\u51c6\u4e3b\u8981\u5206\u4e3a\u4e09\u4e2a\u9886\u57df\uff1a<strong>\u901a\u7528\u8bed\u8a00\u4efb\u52a1<\/strong>\u57fa\u51c6\u6d4b\u8bd5\u3001<strong>\u7279\u5b9a\u4e0b\u6e38\u4efb\u52a1<\/strong>\u57fa\u51c6\u6d4b\u8bd5\u4ee5\u53ca<strong>\u591a\u6a21\u6001\u4efb\u52a1<\/strong>\u57fa\u51c6\u6d4b\u8bd5\u3002<\/p>\n<table>\n<thead>\n<tr>\n<th>\u57fa\u51c6\u6d4b\u8bd5<\/th>\n<th>\u91cd\u70b9\u5173\u6ce8<\/th>\n<th>\u9886\u57df<\/th>\n<th>\u8bc4\u4f30\u6807\u51c6<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>SOCKET [23]<\/td>\n<td>\u793e\u4f1a\u77e5\u8bc6<\/td>\n<td>\u7279\u5b9a\u4e0b\u6e38\u4efb\u52a1<\/td>\n<td>\u793e\u4f1a\u8bed\u8a00\u7406\u89e3\u80fd\u529b<\/td>\n<\/tr>\n<tr>\n<td>MME[46]<\/td>\n<td>\u591a\u6a21\u6001\u5927\u8bed\u8a00\u6a21\u578b<\/td>\n<td>\u591a\u6a21\u6001\u4efb\u52a1<\/td>\n<td>\u611f\u77e5\u4e0e\u8ba4\u77e5\u80fd\u529b<\/td>\n<\/tr>\n<tr>\n<td>\u9e2e\uff08Xiezhi\uff09 [59]<\/td>\n<td>\u7efc\u5408\u9886\u57df\u77e5\u8bc6<\/td>\n<td>\u901a\u7528\u8bed\u8a00\u4efb\u52a1<\/td>\n<td>\u591a\u4e2a\u57fa\u51c6\u6d4b\u8bd5\u7684\u6574\u4f53\u6027\u80fd<\/td>\n<\/tr>\n<tr>\n<td>Choice &#8211; 75[75]<\/td>\n<td>\u811a\u672c\u5b66\u4e60<\/td>\n<td>\u7279\u5b9a\u4e0b\u6e38\u4efb\u52a1<\/td>\n<td>\u5927\u8bed\u8a00\u6a21\u578b\u7684\u6574\u4f53\u6027\u80fd<\/td>\n<\/tr>\n<tr>\n<td>CUAD71<\/td>\n<td>\u6cd5\u5f8b\u5408\u540c\u5ba1\u67e5<\/td>\n<td>\u7279\u5b9a\u4e0b\u6e38\u4efb\u52a1<\/td>\n<td>\u6cd5\u5f8b\u5408\u540c\u7406\u89e3\u80fd\u529b<\/td>\n<\/tr>\n<tr>\n<td>TRUSTGPT[79]<\/td>\n<td>\u4f26\u7406<\/td>\n<td>\u7279\u5b9a\u4e0b\u6e38\u4efb\u52a1<\/td>\n<td>\u6bd2\u6027\u3001\u504f\u5dee\u4e0e\u4ef7\u503c\u4e00\u81f4\u6027<\/td>\n<\/tr>\n<tr>\n<td>MMLU[70]<\/td>\n<td>\u6587\u672c\u6a21\u578b<\/td>\n<td>\u901a\u7528\u8bed\u8a00\u4efb\u52a1<\/td>\n<td>\u591a\u4efb\u52a1\u51c6\u786e\u7387<\/td>\n<\/tr>\n<tr>\n<td>MATH[72]<\/td>\n<td>\u6570\u5b66\u95ee\u9898<\/td>\n<td>\u7279\u5b9a\u4e0b\u6e38\u4efb\u52a1<\/td>\n<td>\u6570\u5b66\u80fd\u529b<\/td>\n<\/tr>\n<tr>\n<td>APPS [68]<\/td>\n<td>\u7f16\u7801\u6311\u6218\u80fd\u529b<\/td>\n<td>\u7279\u5b9a\u4e0b\u6e38\u4efb\u52a1<\/td>\n<td>\u4ee3\u7801\u751f\u6210\u80fd\u529b<\/td>\n<\/tr>\n<tr>\n<td>CELLO[66] <br \/> C &#8211; Eval [78]<\/td>\n<td>\u590d\u6742\u6307\u4ee4 <br \/> \u4e2d\u6587\u8bc4\u4f30<\/td>\n<td>\u7279\u5b9a\u4e0b\u6e38\u4efb\u52a1 <br \/> \u901a\u7528\u8bed\u8a00\u4efb\u52a1<\/td>\n<td>\u56db\u9879\u6307\u5b9a\u8bc4\u4f30\u6807\u51c6 <br \/> \u4e2d\u6587\u8bed\u5883\u4e0b\u768452\u9879\u8003\u8bd5<\/td>\n<\/tr>\n<tr>\n<td>EmotionBench[76]<\/td>\n<td>\u5171\u60c5\u80fd\u529b<\/td>\n<td>\u7279\u5b9a\u4e0b\u6e38\u4efb\u52a1<\/td>\n<td>\u60c5\u7eea\u53d8\u5316<\/td>\n<\/tr>\n<tr>\n<td>OpenLLM[80]<\/td>\n<td>\u804a\u5929\u673a\u5668\u4eba<\/td>\n<td>\u901a\u7528\u8bed\u8a00\u4efb\u52a1<\/td>\n<td>\u6392\u884c\u699c\u6392\u540d<\/td>\n<\/tr>\n<tr>\n<td>DynaBench [94]<\/td>\n<td>\u52a8\u6001\u8bc4\u4f30<\/td>\n<td>\u901a\u7528\u8bed\u8a00\u4efb\u52a1<\/td>\n<td>\u81ea\u7136\u8bed\u8a00\u63a8\u7406\u3001\u95ee\u7b54\u3001\u60c5\u611f\u5206\u6790\u4e0e\u4ec7\u6068\u8a00\u8bba\u68c0\u6d4b<\/td>\n<\/tr>\n<tr>\n<td>Chatbot Arena [128]<\/td>\n<td>\u804a\u5929\u52a9\u624b<\/td>\n<td>\u901a\u7528\u8bed\u8a00\u4efb\u52a1<\/td>\n<td>\u4f17\u5305\u548cElo\u8bc4\u7ea7\u7cfb\u7edf<\/td>\n<\/tr>\n<tr>\n<td>AlpacaEval [112]<\/td>\n<td>\u81ea\u52a8\u8bc4\u4f30<\/td>\n<td>\u901a\u7528\u8bed\u8a00\u4efb\u52a1<\/td>\n<td>\u6307\u6807\u3001\u7a33\u5065\u6027\u4e0e\u591a\u6837\u6027<\/td>\n<\/tr>\n<tr>\n<td>CMMLU[108]<\/td>\n<td>\u4e2d\u6587\u591a\u4efb\u52a1\u5904\u7406<\/td>\n<td>\u7279\u5b9a\u4e0b\u6e38\u4efb\u52a1<\/td>\n<td>\u591a\u4efb\u52a1\u8bed\u8a00\u7406\u89e3\u80fd\u529b<\/td>\n<\/tr>\n<tr>\n<td>HELM[114]<\/td>\n<td>\u6574\u4f53\u8bc4\u4f30<\/td>\n<td>\u901a\u7528\u8bed\u8a00\u4efb\u52a1<\/td>\n<td>\u591a\u6307\u6807<\/td>\n<\/tr>\n<tr>\n<td>API &#8211; Bank [109]<\/td>\n<td>\u5de5\u5177\u5229\u7528<\/td>\n<td>\u7279\u5b9a\u4e0b\u6e38\u4efb\u52a1<\/td>\n<td>API\u8c03\u7528\u3001\u68c0\u7d22\u4e0e\u89c4\u5212\u80fd\u529b<\/td>\n<\/tr>\n<tr>\n<td>M3KE[122]<\/td>\n<td>\u591a\u4efb\u52a1<\/td>\n<td>\u7279\u5b9a\u4e0b\u6e38\u4efb\u52a1<\/td>\n<td>\u591a\u4efb\u52a1\u51c6\u786e\u7387<\/td>\n<\/tr>\n<tr>\n<td>MMBench[126]<\/td>\n<td>\u5927\u578b\u89c6\u89c9 &#8211; \u8bed\u8a00\u6a21\u578b\uff08LVLMs\uff09<\/td>\n<td>\u591a\u6a21\u6001\u4efb\u52a1<\/td>\n<td>\u89c6\u89c9 &#8211; \u8bed\u8a00\u6a21\u578b\u7684\u591a\u65b9\u9762\u80fd\u529b<\/td>\n<\/tr>\n<tr>\n<td>SEED &#8211; Bench [107]<\/td>\n<td>\u591a\u6a21\u6001\u5927\u8bed\u8a00\u6a21\u578b<\/td>\n<td>\u591a\u6a21\u6001\u4efb\u52a1<\/td>\n<td>\u591a\u6a21\u6001\u5927\u8bed\u8a00\u6a21\u578b\u7684\u751f\u6210\u6027\u7406\u89e3\u80fd\u529b<\/td>\n<\/tr>\n<tr>\n<td>UHGEval [116]<\/td>\n<td>\u4e2d\u6587\u5927\u8bed\u8a00\u6a21\u578b\u7684\u5e7b\u89c9\u95ee\u9898<\/td>\n<td>\u7279\u5b9a\u4e0b\u6e38\u4efb\u52a1<\/td>\n<td>\u5f62\u5f0f\u3001\u6307\u6807\u4e0e\u7c92\u5ea6<\/td>\n<\/tr>\n<tr>\n<td>ARB[171]<\/td>\n<td>\u9ad8\u7ea7\u63a8\u7406\u80fd\u529b<\/td>\n<td>\u7279\u5b9a\u4e0b\u6e38\u4efb\u52a1<\/td>\n<td>\u591a\u9886\u57df\u9ad8\u7ea7\u63a8\u7406\u80fd\u529b<\/td>\n<\/tr>\n<tr>\n<td>BIG &#8211; bench [182]<\/td>\n<td>\u5927\u8bed\u8a00\u6a21\u578b\u7684\u80fd\u529b\u4e0e\u5c40\u9650<\/td>\n<td>\u901a\u7528\u8bed\u8a00\u4efb\u52a1<\/td>\n<td>\u6a21\u578b\u6027\u80fd\u4e0e\u6821\u51c6<\/td>\n<\/tr>\n<tr>\n<td>MultiMedQA[177]<\/td>\n<td>\u533b\u5b66\u95ee\u7b54<\/td>\n<td>\u7279\u5b9a\u4e0b\u6e38\u4efb\u52a1<\/td>\n<td>\u51c6\u786e\u7387\u4e0e\u4eba\u8bc4<\/td>\n<\/tr>\n<tr>\n<td>CVALUES[230]<\/td>\n<td>\u5b89\u5168\u6027\u4e0e\u8d23\u4efb\u6027<\/td>\n<td>\u7279\u5b9a\u4e0b\u6e38\u4efb\u52a1<\/td>\n<td>\u5927\u8bed\u8a00\u6a21\u578b\u7684\u5bf9\u9f50\u80fd\u529b<\/td>\n<\/tr>\n<tr>\n<td>LVLM &#8211; eHub[231]<\/td>\n<td>\u5927\u578b\u89c6\u89c9 &#8211; \u8bed\u8a00\u6a21\u578b<\/td>\n<td>\u591a\u6a21\u6001\u4efb\u52a1<\/td>\n<td>\u5927\u578b\u89c6\u89c9 &#8211; \u8bed\u8a00\u6a21\u578b\u7684\u591a\u6a21\u6001\u80fd\u529b<\/td>\n<\/tr>\n<tr>\n<td>ToolBench[191]<\/td>\n<td>\u8f6f\u4ef6\u5de5\u5177<\/td>\n<td>\u7279\u5b9a\u4e0b\u6e38\u4efb\u52a1<\/td>\n<td>\u6267\u884c\u6210\u529f\u7387<\/td>\n<\/tr>\n<tr>\n<td>FRESHQA[198]<\/td>\n<td>\u52a8\u6001\u95ee\u7b54<\/td>\n<td>\u7279\u5b9a\u4e0b\u6e38\u4efb\u52a1<\/td>\n<td>\u6b63\u786e\u6027\u4e0e\u5e7b\u89c9\u95ee\u9898<\/td>\n<\/tr>\n<tr>\n<td>CMB[211]<\/td>\n<td>\u4e2d\u533b\u7efc\u5408<\/td>\n<td>\u7279\u5b9a\u4e0b\u6e38\u4efb\u52a1<\/td>\n<td>\u4e13\u5bb6\u8bc4\u4f30\u4e0e\u81ea\u52a8\u8bc4\u4f30<\/td>\n<\/tr>\n<tr>\n<td>PandaLM[216]<\/td>\n<td>\u6307\u4ee4\u5fae\u8c03<\/td>\n<td>\u901a\u7528\u8bed\u8a00\u4efb\u52a1<\/td>\n<td>\u7531PandaLM\u5224\u65ad\u7684\u80dc\u7387<\/td>\n<\/tr>\n<tr>\n<td>MINT [213]<\/td>\n<td>\u591a\u8f6e\u4ea4\u4e92<\/td>\n<td>\u7279\u5b9a\u4e0b\u6e38\u4efb\u52a1<\/td>\n<td>k\u8f6e\u9884\u7b97\u6210\u529f\u7387SRk<\/td>\n<\/tr>\n<tr>\n<td>Dialogue CoT[205]<\/td>\n<td>\u6df1\u5ea6\u5bf9\u8bdd<\/td>\n<td>\u7279\u5b9a\u4e0b\u6e38\u4efb\u52a1<\/td>\n<td>\u5927\u8bed\u8a00\u6a21\u578b\u7684\u6709\u7528\u6027\u4e0e\u53ef\u63a5\u53d7\u6027<\/td>\n<\/tr>\n<tr>\n<td>BOSS[239]<\/td>\n<td>\u81ea\u7136\u8bed\u8a00\u5904\u7406\u4e2d\u7684\u5206\u5e03\u5916\u7a33\u5065\u6027<\/td>\n<td>\u901a\u7528\u8bed\u8a00\u4efb\u52a1<\/td>\n<td>\u5206\u5e03\u5916\u7a33\u5065\u6027<\/td>\n<\/tr>\n<tr>\n<td>MM &#8211; Vet [238]<\/td>\n<td>\u590d\u6742\u591a\u6a21\u6001\u4efb\u52a1<\/td>\n<td>\u591a\u6a21\u6001\u4efb\u52a1<\/td>\n<td>\u7efc\u5408\u89c6\u89c9 &#8211; \u8bed\u8a00\u80fd\u529b<\/td>\n<\/tr>\n<tr>\n<td>LAMM[235]<\/td>\n<td>\u591a\u6a21\u6001\u70b9\u4e91<\/td>\n<td>\u591a\u6a21\u6001\u4efb\u52a1<\/td>\n<td>\u7279\u5b9a\u4efb\u52a1\u6307\u6807<\/td>\n<\/tr>\n<tr>\n<td>GLUE &#8211; X[234]<\/td>\n<td>\u81ea\u7136\u8bed\u8a00\u5904\u7406\u4efb\u52a1\u7684\u5206\u5e03\u5916\u7a33\u5065\u6027<\/td>\n<td>\u901a\u7528\u8bed\u8a00\u4efb\u52a1<\/td>\n<td>\u5206\u5e03\u5916\u7a33\u5065\u6027<\/td>\n<\/tr>\n<tr>\n<td>KoLA[236]<\/td>\n<td>\u77e5\u8bc6\u5bfc\u5411\u8bc4\u4f30<\/td>\n<td>\u901a\u7528\u8bed\u8a00\u4efb\u52a1<\/td>\n<td>\u81ea\u5bf9\u6bd4\u6307\u6807<\/td>\n<\/tr>\n<tr>\n<td>AGIEval [262]<\/td>\n<td>\u4ee5\u4eba\u4e3a\u4e2d\u5fc3\u7684\u57fa\u7840\u6a21\u578b<\/td>\n<td>\u901a\u7528\u8bed\u8a00\u4efb\u52a1<\/td>\n<td>\u901a\u7528\u6307\u6807<\/td>\n<\/tr>\n<tr>\n<td>PromptBench [264]<\/td>\n<td>\u5bf9\u6297\u6027\u63d0\u793a\u6297\u6027<\/td>\n<td>\u901a\u7528\u8bed\u8a00\u4efb\u52a1<\/td>\n<td>\u5bf9\u6297\u7a33\u5065\u6027<\/td>\n<\/tr>\n<tr>\n<td>MT &#8211; Bench [260]<\/td>\n<td>\u591a\u8f6e\u5bf9\u8bdd<\/td>\n<td>\u901a\u7528\u8bed\u8a00\u4efb\u52a1<\/td>\n<td>\u7531GPT &#8211; 4\u5224\u65ad\u7684\u80dc\u7387<\/td>\n<\/tr>\n<tr>\n<td>M3Exam [250]<\/td>\n<td>\u591a\u8bed\u8a00\u3001\u591a\u6a21\u6001\u4e0e\u591a\u5c42\u6b21<\/td>\n<td>\u7279\u5b9a\u4e0b\u6e38\u4efb\u52a1<\/td>\n<td>\u7279\u5b9a\u4efb\u52a1\u6307\u6807<\/td>\n<\/tr>\n<tr>\n<td>GAOKAO &#8211; Bench245<\/td>\n<td>\u4e2d\u56fd\u9ad8\u8003\u8003\u8bd5<\/td>\n<td>\u7279\u5b9a\u4e0b\u6e38\u4efb\u52a1<\/td>\n<td>\u51c6\u786e\u7387\u4e0e\u5f97\u5206\u7387<\/td>\n<\/tr>\n<tr>\n<td>SafetyBench [254]<\/td>\n<td>\u5b89\u5168\u6027<\/td>\n<td>\u7279\u5b9a\u4e0b\u6e38\u4efb\u52a1<\/td>\n<td>\u5927\u8bed\u8a00\u6a21\u578b\u7684\u5b89\u5168\u80fd\u529b<\/td>\n<\/tr>\n<tr>\n<td>LLMEval [252]<\/td>\n<td>\u5927\u8bed\u8a00\u6a21\u578b\u8bc4\u4f30\u5668<\/td>\n<td>\u901a\u7528\u8bed\u8a00\u4efb\u52a1<\/td>\n<td>\u51c6\u786e\u7387\u3001\u5b8fF1\u503c\u548ckappa\u76f8\u5173\u7cfb\u6570<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>\u4e3e\u4f8b\u8bf4\u660e\uff1a<\/p>\n<ul>\n<li>\n<p>MATH\u57fa\u51c6\uff1a<\/p>\n<ul>\n<li>\u8be5\u57fa\u51c6\u6d4b\u8bd5\u5305\u542b12,500\u4e2a\u6570\u5b66\u95ee\u9898\uff0c\u6db5\u76d6\u51e0\u4f55\u3001\u4ee3\u6570\u3001\u6570\u8bba\u7b497\u4e2a\u9886\u57df\uff0c\u9898\u578b\u5305\u62ec\u9009\u62e9\u9898\u548c\u8bc1\u660e\u9898\uff0c\u96be\u5ea6\u4ece\u521d\u4e2d\u5230\u56fd\u9645\u6570\u5b66\u5965\u6797\u5339\u514b\u7ade\u8d5b\u7ea7\u522b\u3002<\/li>\n<li>\u6d4b\u8bd5\u6570\u636e\u96c6\u4ed3\u5e93\u5730\u5740\uff1a<a href=\"https:\/\/github.com\/hendrycks\/math\">https:\/\/github.com\/hendrycks\/math<\/a><\/li>\n<\/ul>\n<\/li>\n<li>\n<p>APPS\u57fa\u51c6\uff1a<\/p>\n<ul>\n<li>\u8be5\u57fa\u51c6\u6d4b\u8bd5\u6536\u96c6\u4e8610,000\u4e2a\u7f16\u7a0b\u9898\u76ee\uff0c\u96be\u5ea6\u5bf9\u6807LeetCode\u4e2d\u7b49\u4ee5\u4e0a\u96be\u5ea6\uff0c\u5305\u542b\u7b97\u6cd5\u8bbe\u8ba1\u3001\u8fb9\u754c\u6761\u4ef6\u5904\u7406\u7b49\u6d4b\u8bd5\u7528\u4f8b\u3002<\/li>\n<li>\u6d4b\u8bd5\u6570\u636e\u96c6\u4ed3\u5e93\u5730\u5740\uff1a<a href=\"https:\/\/github.com\/hendrycks\/apps\">https:\/\/github.com\/hendrycks\/apps<\/a><\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<blockquote>\n<p>\u5907\u6ce8\uff1a<br \/>\n\u4ee5\u4e0a\u57fa\u51c6\u6d4b\u8bd5\u7684\u90e8\u5206\u5185\u5bb9\u7531deepseek\u8f85\u52a9\u751f\u6210\uff0c\u672c\u4eba\u4ec5\u5bf9\u6570\u636e\u96c6\u4ed3\u5e93\u5730\u5740\u8fdb\u884c\u4e86\u6c42\u8bc1\uff0c\u5176\u4ed6\u4fe1\u606f\u5e76\u672a\u6df1\u5165\u8003\u7a76\u3002<\/p>\n<\/blockquote>\n<h2><span class=\"ez-toc-section\" id=\"How%EF%BC%9A%E5%A6%82%E4%BD%95%E8%AF%84%E4%BB%B7\"><\/span>How\uff1a\u5982\u4f55\u8bc4\u4ef7<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>\u5927\u6a21\u578b\u8bc4\u4f30\u4e0e\u4f20\u7edf\u8f6f\u4ef6\u7684\u8bc4\u6d4b\u601d\u60f3\u4e00\u81f4\uff0c\u91c7\u7528\u5ba2\u89c2\u8bc4\u4ef7(\u81ea\u52a8\u8bc4\u4f30)\u548c\u4e3b\u89c2\u8bc4\u4ef7(\u4eba\u5de5\u8bc4\u4f30)\u76f8\u7ed3\u5408\u7684\u8bc4\u4ef7\u65b9\u5f0f\uff0c\u5177\u4f53\u5c55\u5f00\u5185\u5bb9\u5982\u4e0b\uff1a<\/p>\n<h3><span class=\"ez-toc-section\" id=\"1_%E8%87%AA%E5%8A%A8%E8%AF%84%E4%BC%B0%EF%BC%88Automatic_Evaluation%EF%BC%89\"><\/span>1. \u81ea\u52a8\u8bc4\u4f30\uff08Automatic Evaluation\uff09<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p><strong>\u6838\u5fc3\u7279\u5f81<\/strong>\uff1a<\/p>\n<ul>\n<li>\u65e0\u9700\u4eba\u5de5\u53c2\u4e0e\uff0c\u901a\u8fc7\u9884\u5b9a\u4e49\u6307\u6807\u91cf\u5316\u8bc4\u4f30<\/li>\n<li>\u8bc4\u4f30\u8fc7\u7a0b\u6807\u51c6\u5316\u3001\u53ef\u91cd\u590d<\/li>\n<\/ul>\n<p><strong>\u5178\u578b\u6307\u6807<\/strong>\uff1a<\/p>\n<table>\n<thead>\n<tr>\n<th>\u8bc4\u4f30\u7ef4\u5ea6<\/th>\n<th>\u6307\u6807<\/th>\n<th>\u8ba1\u7b97\u516c\u5f0f<\/th>\n<th>\u5e94\u7528\u573a\u666f\u793a\u4f8b<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>\u51c6\u786e\u6027<\/td>\n<td>\u7cbe\u786e\u5339\u914d(EM)<\/td>\n<td><code class=\"katex-inline\">\\text{EM} = \\frac{\\sum \\mathbb{I}(pred=ref)}{N}<\/code><\/td>\n<td>\u95ed\u5377\u95ee\u7b54\u3001\u4ee3\u7801\u751f\u6210<\/td>\n<\/tr>\n<tr>\n<td><\/td>\n<td>F1 Score<\/td>\n<td><code class=\"katex-inline\">F1 = \\frac{2 \\times P \\times R}{P + R}<\/code><\/td>\n<td>\u6587\u672c\u5206\u7c7b\u3001\u5b9e\u4f53\u8bc6\u522b<\/td>\n<\/tr>\n<tr>\n<td><\/td>\n<td>ROUGE-L<\/td>\n<td>\u6682\u7565<\/td>\n<td>\u6458\u8981\u751f\u6210\u3001\u673a\u5668\u7ffb\u8bd1<\/td>\n<\/tr>\n<tr>\n<td>\u6821\u51c6\u5ea6<\/td>\n<td>\u671f\u671b\u6821\u51c6\u8bef\u5dee(ECE)<\/td>\n<td>\u6682\u7565<\/td>\n<td>\u533b\u7597\u8bca\u65ad\u3001\u98ce\u9669\u8bc4\u4f30<\/td>\n<\/tr>\n<tr>\n<td>\u516c\u5e73\u6027<\/td>\n<td>\u4eba\u53e3\u5747\u7b49\u5dee\u5f02(DPD)<\/td>\n<td><code class=\"katex-inline\">DPD = P(\\hat{y}\\|Z=1) - P(\\hat{y}\\|Z=0)<\/code><\/td>\n<td>\u62db\u8058\u6587\u6848\u751f\u6210\u3001\u4fe1\u7528\u8bc4\u4f30<\/td>\n<\/tr>\n<tr>\n<td>\u9c81\u68d2\u6027<\/td>\n<td>\u653b\u51fb\u6210\u529f\u7387(ASR)<\/td>\n<td><code class=\"katex-inline\">ASR = \\frac{\\sum \\mathbb{I}(f(A(x)) \\neq y)}{\\sum \\mathbb{I}(f(x)=y)}<\/code><\/td>\n<td>\u5bf9\u6297\u653b\u51fb\u6d4b\u8bd5\u3001\u8f93\u5165\u6270\u52a8\u6d4b\u8bd5<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p><strong>\u5de5\u5177\u751f\u6001<\/strong>\uff1a<\/p>\n<table>\n<thead>\n<tr>\n<th>\u8bc4\u6d4b\u5de5\u5177<\/th>\n<th>\u5de5\u5177\u94fe\u63a5<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>lm &#8211; evaluation &#8211; harness<\/td>\n<td><a href=\"https:\/\/github.com\/EleutherAI\/lm-evaluation-harness\">https:\/\/github.com\/EleutherAI\/lm-evaluation-harness<\/a><\/td>\n<\/tr>\n<tr>\n<td>OpenCompasss<\/td>\n<td><a href=\"https:\/\/opencompass.org.cn\/\">https:\/\/opencompass.org.cn\/<\/a><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>(\u5f85\u8865\u5145\u5b8c\u5584)<\/p>\n<h3><span class=\"ez-toc-section\" id=\"2_%E4%BA%BA%E5%B7%A5%E8%AF%84%E4%BC%B0%EF%BC%88Human_Evaluation%EF%BC%89\"><\/span>2. \u4eba\u5de5\u8bc4\u4f30\uff08Human Evaluation\uff09<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p><strong>\u8bc4\u4f30\u6846\u67b6<\/strong>\uff1a<\/p>\n<table>\n<thead>\n<tr>\n<th>\u5173\u952e\u8981\u7d20<\/th>\n<th>\u8981\u6c42\u8bf4\u660e<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>\u8bc4\u4f30\u8005\u6570\u91cf<\/td>\n<td>\u6bcf\u7ec4\u22653\u4eba\uff0c\u4fdd\u8bc1\u7edf\u8ba1\u663e\u8457\u6027<\/td>\n<\/tr>\n<tr>\n<td>\u8bc4\u4f30\u6807\u51c6<\/td>\n<td>\u51c6\u786e\u6027\u3001\u76f8\u5173\u6027\u3001\u6d41\u7545\u6027\u3001\u5b89\u5168\u6027\u3001\u900f\u660e\u5ea6\u3001\u5b89\u5168\u6027\u3001\u4eba\u7c7b\u4e00\u81f4\u6027\u7b49<\/td>\n<\/tr>\n<tr>\n<td>\u8bc4\u4f30\u8005\u8d44\u8d28<\/td>\n<td>\u9886\u57df\u4e13\u5bb6\u5360\u6bd4\u226530%\uff0c\u5747\u9700\u901a\u8fc7\u8bc4\u4f30\u57f9\u8bad<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p><strong>\u5b9e\u65bd\u6d41\u7a0b<\/strong>\uff1a<\/p>\n<ol>\n<li><strong>\u8bbe\u8ba1\u8bc4\u4f30\u77e9\u9635<\/strong>\uff1a\n<pre><code class=\"language-python\"># \u8bc4\u4f30\u7ef4\u5ea6\u6743\u91cd\u914d\u7f6e\u793a\u4f8b\ncriteria = {\n   '\u51c6\u786e\u6027': 0.3,\n   '\u76f8\u5173\u6027': 0.2,\n   '\u6d41\u7545\u6027': 0.15,\n   '\u5b89\u5168\u6027': 0.2,\n   '\u900f\u660e\u5ea6': 0.15\n}<\/code><\/pre>\n<\/li>\n<li><strong>\u6267\u884c\u53cc\u76f2\u8bc4\u4f30<\/strong>\uff1a\u8bc4\u4f30\u8005\u4e0d\u77e5\u6653\u6a21\u578b\u7248\u672c\u4fe1\u606f<\/li>\n<li><strong>\u7edf\u8ba1\u5206\u6790<\/strong>\uff1a\u4f7f\u7528Krippendorff&#8217;s alpha\u7cfb\u6570\u8ba1\u7b97\u8bc4\u5206\u8005\u95f4\u4fe1\u5ea6<\/li>\n<\/ol>\n<p>\u8bc4\u4f30\u65b9\u6cd5\u5bf9\u6bd4<\/p>\n<table>\n<thead>\n<tr>\n<th>\u7ef4\u5ea6<\/th>\n<th>\u81ea\u52a8\u8bc4\u4f30<\/th>\n<th>\u4eba\u5de5\u8bc4\u4f30<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td><strong>\u6267\u884c\u6210\u672c<\/strong><\/td>\n<td>\u4f4e<\/td>\n<td>\u9ad8<\/td>\n<\/tr>\n<tr>\n<td><strong>\u8bc4\u4f30\u5468\u671f<\/strong><\/td>\n<td>\u5206\u949f\u7ea7<\/td>\n<td>\u5929\u7ea7<\/td>\n<\/tr>\n<tr>\n<td><strong>\u53ef\u89e3\u91ca\u6027<\/strong><\/td>\n<td>\u91cf\u5316\u7ed3\u679c\u660e\u786e\u4f46\u53ef\u89e3\u91ca\u6027\u5dee<\/td>\n<td>\u53ef\u63d0\u4f9b\u5b9a\u6027\u53cd\u9988<\/td>\n<\/tr>\n<tr>\n<td><strong>\u9002\u7528\u8303\u56f4<\/strong><\/td>\n<td>\u6807\u51c6\u5316\u4efb\u52a1\uff08\u5206\u7c7b\u3001\u751f\u6210\u7b49\uff09<\/td>\n<td>\u521b\u9020\u6027\u4efb\u52a1\uff08\u5199\u4f5c\u3001\u8bbe\u8ba1\u7b49\uff09<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<h3><span class=\"ez-toc-section\" id=\"3_%E5%89%8D%E6%B2%BF%E8%AF%84%E4%BC%B0%E6%96%B9%E6%B3%95%E6%8E%A2%E7%B4%A2\"><\/span>3. \u524d\u6cbf\u8bc4\u4f30\u65b9\u6cd5\u63a2\u7d22<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p>\u9664\u4e86\u4e0a\u8ff0\u4e24\u79cd\u8bc4\u4f30\u65b9\u5f0f\u4e4b\u5916\uff0c\u73b0\u5728\u8fd8\u51fa\u73b0\u4e86\u4e00\u4e9b\u524d\u6cbf\u7684\u8bc4\u4f30\u65b9\u6cd5\uff0c\u4f8b\u5982\uff1a<\/p>\n<p><strong>\u601d\u7ef4\u94fe\u8bc4\u4f30\uff08CoT Evaluation\uff09<\/strong>\uff1a<\/p>\n<pre><code class=\"language-python\"># \u4f7f\u7528GPT-4\u8fdb\u884c\u81ea\u52a8\u8bc4\u4f30\ndef cot_evaluation(prompt, response):\n    evaluation_prompt = f&quot;&quot;&quot;\n    \u8bf7\u8bc4\u4f30\u4ee5\u4e0b\u56de\u7b54\u7684\u8d28\u91cf\uff081-5\u5206\uff09\uff1a\n    \u95ee\u9898\uff1a{prompt}\n    \u56de\u7b54\uff1a{response}\n    \u8bc4\u4f30\u6807\u51c6\uff1a\n    1. \u4e8b\u5b9e\u51c6\u786e\u6027 2. \u903b\u8f91\u8fde\u8d2f\u6027 3. \u6f5c\u5728\u5371\u5bb3\u6027\n    &quot;&quot;&quot;\n    return gpt4_api(evaluation_prompt)\n\n# \u6267\u884c\u6279\u91cf\u8bc4\u4f30\nscores = [cot_evaluation(p, r) for p, r in zip(prompts, responses)]<\/code><\/pre>\n<p><strong>\u591a\u6a21\u6001\u8bc4\u4f30\u6846\u67b6<\/strong>\uff1a<\/p>\n<pre><code class=\"language-mermaid\">graph TD\n    A[\u8f93\u5165] --&gt; B[\u6587\u672c\u5206\u6790]\n    A --&gt; C[\u56fe\u50cf\u8bc6\u522b]\n    A --&gt; D[\u8bed\u97f3\u5904\u7406]\n    B --&gt; E[\u8bed\u4e49\u7406\u89e3\u8bc4\u5206]\n    C --&gt; F[\u89c6\u89c9\u4e00\u81f4\u6027\u8bc4\u5206]\n    D --&gt; G[\u8bed\u97f3\u81ea\u7136\u5ea6\u8bc4\u5206]\n    E --&gt; H[\u7efc\u5408\u8bc4\u4f30]\n    F --&gt; H\n    G --&gt; H<\/code><\/pre>\n<blockquote>\n<p>\u8bba\u6587\u542f\u793a\uff1a\u6700\u65b0\u7814\u7a76\u663e\u793a\uff0c\u7ed3\u5408\u81ea\u52a8\u8bc4\u4f30\u7684\u6548\u7387\u4e0e\u4eba\u5de5\u8bc4\u4f30\u7684\u6df1\u5ea6\uff0c\u91c7\u7528\u300cAI-Human Hybrid\u300d\u6a21\u5f0f\u53ef\u83b7\u5f97\u6700\u4f18\u8bc4\u4f30\u6548\u679c\uff08Bubeck et al., 2023\uff09<\/p>\n<\/blockquote>\n<h2><span class=\"ez-toc-section\" id=\"%E5%86%85%E5%AE%B9%E5%B0%8F%E7%BB%93\"><\/span>\u5185\u5bb9\u5c0f\u7ed3<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<ul>\n<li>\n<p>\u5927\u6a21\u578b\u8bc4\u6d4b\u81f3\u5173\u91cd\u8981\uff1a<\/p>\n<ul>\n<li><strong>\u5b83\u6709\u52a9\u4e8e\u6211\u4eec\u66f4\u597d\u5730\u4e86\u89e3\u5927\u6a21\u578b\u4f18\u52bf\u548c\u52a3\u52bf<\/strong>\u3002 <\/li>\n<li>\u53ef\u4ee5\u66f4\u597d\u5730\u4e3a<strong>\u4eba\u7c7b\u4e0e\u5927\u6a21\u578b\u7684\u4ea4\u4e92\u63d0\u4f9b\u6307\u5bfc<\/strong>\u3002<\/li>\n<li>\u66f4\u597d\u5730<strong>\u7edf\u7b79\u548c\u89c4\u5212\u5927\u6a21\u578b\u672a\u6765\u7684\u53d1\u5c55<\/strong>\u548c\u6f14\u53d8\u3002<\/li>\n<\/ul>\n<\/li>\n<li>\n<p>\u5927\u6a21\u578b\u8bc4\u6d4b\u7684\u8bc4\u4ef7\u4f53\u7cfb<\/p>\n<ul>\n<li>\u8bc4\u4ef7\u4f53\u7cfb\u9700\u8981\u5305\u542b<strong>\u53ef\u9760\u6027<\/strong>\u3001<strong>\u5b89\u5168\u6027<\/strong>\u3001<strong>\u516c\u5e73\u6027<\/strong>\u3001<strong>\u62b5\u5fa1\u6ee5\u7528\u80fd\u529b<\/strong>\u3001<strong>\u53ef\u89e3\u91ca\u6027\u4e0e\u63a8\u7406\u80fd\u529b<\/strong>\u3001<strong>\u793e\u4f1a\u89c4\u8303<\/strong>\u3001<strong>\u9c81\u68d2\u6027<\/strong>\u7b49\u7ef4\u5ea6\u3002<\/li>\n<li>\u8bc4\u4ef7\u4f53\u7cfb\u9700\u8981\u5305\u542b<strong>\u901a\u7528\u8bed\u8a00\u4efb\u52a1<\/strong>\u3001<strong>\u7279\u5b9a\u4e0b\u6e38\u4efb\u52a1<\/strong>\u3001<strong>\u591a\u6a21\u6001\u4efb\u52a1<\/strong>\u7b49\u9886\u57df\u7684\u8bc4\u4ef7\u3002<\/li>\n<\/ul>\n<\/li>\n<li>\n<p>\u5927\u6a21\u578b\u8bc4\u6d4b\u7684\u8bc4\u4ef7\u65b9\u6cd5<\/p>\n<ul>\n<li>\u8bc4\u4ef7\u65b9\u6cd5\u9700\u8981\u5305\u542b<strong>\u81ea\u52a8\u8bc4\u4f30<\/strong>\u3001<strong>\u4eba\u5de5\u8bc4\u4f30<\/strong>\u4e24\u79cd\u65b9\u6cd5\u3002<\/li>\n<li>\u81ea\u52a8\u8bc4\u4f30\u501f\u52a9\u5de5\u5177\u8fdb\u884c\u81ea\u52a8\u5316\u8bc4\u4f30\uff0c\u4e3b\u8981\u8bc4\u4f30\u7684\u6307\u6807\u6709\uff1a\u7cbe\u786e\u5339\u914d(EM)\u3001F1 Score\u3001ROUGE-L\u3001\u6821\u51c6\u5ea6\u3001\u516c\u5e73\u6027\u3001\u9c81\u68d2\u6027\u7b49\u3002<\/li>\n<li>\u4eba\u5de5\u8bc4\u4f30\u9700\u8981\u501f\u52a9\u4eba\u5de5\u8fdb\u884c\u8bc4\u4f30\uff0c\u4e3b\u8981\u8bc4\u4f30\u7684\u6307\u6807\u6709\uff1a\u51c6\u786e\u6027\u3001\u76f8\u5173\u6027\u3001\u6d41\u7545\u6027\u3001\u5b89\u5168\u6027\u3001\u900f\u660e\u5ea6\u3001\u5b89\u5168\u6027\u3001\u4eba\u7c7b\u4e00\u81f4\u6027\u7b49\u3002<\/li>\n<li>\u9664\u4e86\u4e0a\u8ff0\u4e24\u79cd\u8bc4\u4f30\u65b9\u6cd5\u4e4b\u5916\uff0c\u73b0\u5728\u8fd8\u51fa\u73b0\u4e86\u4e00\u4e9b\u524d\u6cbf\u7684\u8bc4\u4f30\u65b9\u6cd5\uff0c\u4f8b\u5982\uff1a\u601d\u7ef4\u94fe\u8bc4\u4f30\u3001\u591a\u6a21\u6001\u8bc4\u4f30\u7b49\u3002<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<h2><span class=\"ez-toc-section\" id=\"%E5%8F%82%E8%80%83%E8%B5%84%E6%96%99\"><\/span>\u53c2\u8003\u8d44\u6599<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<ul>\n<li><a href=\"https:\/\/arxiv.org\/pdf\/2307.03109\">\u300aA Survey on the Evaluation of Large Language Models\u300b<\/a><\/li>\n<li><a href=\"https:\/\/arxiv.org\/pdf\/2308.05374\">\u300aTrustworthy LLMs: a Survey and Guideline for Evaluating Large Language Models\u2019 Alignment\u300b<\/a><\/li>\n<li><a href=\"https:\/\/zhuanlan.zhihu.com\/p\/642689101\">\u77e5\u4e4e\uff1a\u201c\u8bc4\u6d4b\u5373\u79d1\u5b66\u201d\uff1a\u9996\u7bc7\u5927\u8bed\u8a00\u6a21\u578b\u8bc4\u6d4b\u7684\u7efc\u8ff0\uff0c\u4e00\u6587\u5e26\u4f60\u5168\u9762\u4e86\u89e3\u5927\u6a21\u578b\u8bc4\u6d4b\u7684\u73b0\u72b6\u3001\u65b9\u6cd5\u548c\u6311\u6218<\/a><\/li>\n<\/ul>\n<h2><span class=\"ez-toc-section\" id=\"%E9%99%84%E5%BD%95\"><\/span>\u9644\u5f55<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p><a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2025\/02\/\u8bba\u6587\u5185\u5bb9\u5f15\u7528\u7684\u601d\u7ef4\u5bfc\u56fe.jpeg\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2025\/02\/\u8bba\u6587\u5185\u5bb9\u5f15\u7528\u7684\u601d\u7ef4\u5bfc\u56fe.jpeg\" alt=\"\" \/><\/a><\/p>\n<h2><span class=\"ez-toc-section\" id=\"%E5%85%B6%E4%BB%96%E6%96%87%E7%AB%A0\"><\/span>\u5176\u4ed6\u6587\u7ae0<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<ul>\n<li><a href=\"https:\/\/17aitech.com\/?p=38434\">\u3010\u6a21\u578b\u6d4b\u8bd5\u3011\u5927\u6a21\u578b\u6d4b\u8bc4\u4f53\u7cfb\u7684\u6784\u6210<\/a><\/li>\n<li><a href=\"https:\/\/17aitech.com\/?p=38565\">\u3010\u6a21\u578b\u6d4b\u8bd5\u3011\u5927\u6a21\u578b\u8bc4\u6d4b\u5de5\u5177lm-evaluation-harness\u7684\u4f7f\u7528\u65b9\u6cd5\u603b\u7ed3<\/a><\/li>\n<li><a href=\"https:\/\/17aitech.com\/?p=39446\">\u3010\u6a21\u578b\u6d4b\u8bd5\u3011\u5927\u6a21\u578b\u8bc4\u6d4b\u5de5\u5177OpenCompass\u4f7f\u7528\u65b9\u6cd5\u603b\u7ed3<\/a><\/li>\n<li><a href=\"https:\/\/17aitech.com\/?p=39926\">\u3010\u6a21\u578b\u6d4b\u8bd5\u3011ai-eval-system\u5728\u7ebf\u8bc4\u6d4b\u7cfb\u7edfv0.2\u9884\u89c8\u7248\u672c\u4ecb\u7ecd<\/a><\/li>\n<\/ul>\n<p align=\"center\">\u6b22\u8fce\u5173\u6ce8\u516c\u4f17\u53f7\u4ee5\u83b7\u5f97\u6700\u65b0\u7684\u6587\u7ae0\u548c\u65b0\u95fb<\/p>\n<p><a href=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/\u626b\u7801_\u641c\u7d22\u8054\u5408\u4f20\u64ad\u6837\u5f0f-\u767d\u8272\u7248.bmp\" data-fancybox=\"images\" data-fancybox=\"gallery\"><img decoding=\"async\" src=\"https:\/\/17aitech.com\/wp-content\/uploads\/2024\/09\/\u626b\u7801_\u641c\u7d22\u8054\u5408\u4f20\u64ad\u6837\u5f0f-\u767d\u8272\u7248.bmp\" alt=\"\" \/><\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>\u524d\u8a00 \u968f\u7740\u8fd1\u671fDeepsSeek\u5927\u6a21\u578b\u5728 [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":38435,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"aside","meta":{"site-sidebar-layout":"default","site-content-layout":"","ast-site-content-layout":"default","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":"default","adv-header-id-meta":"","stick-header-meta":"default","header-above-stick-meta":"","header-main-stick-meta":"","header-below-stick-meta":"","astra-migrate-meta-layouts":"set","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":[28],"tags":[68,83],"class_list":["post-38434","post","type-post","status-publish","format-aside","has-post-thumbnail","hentry","category-blog","tag-68","tag-83","post_format-post-format-aside"],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v27.9 - https:\/\/yoast.com\/product\/yoast-seo-wordpress\/ -->\n<title>\u3010\u6a21\u578b\u6d4b\u8bd5\u3011\u5927\u6a21\u578b\u6d4b\u8bc4\u4f53\u7cfb\u7684\u6784\u6210 - \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=38434\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\\\/\\\/17aitech.com\\\/?p=38434#article\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/17aitech.com\\\/?p=38434\"},\"author\":{\"name\":\"Dongming\",\"@id\":\"https:\\\/\\\/17aitech.com\\\/#\\\/schema\\\/person\\\/3d23bb6f7f115fcefc9ae7803a691739\"},\"headline\":\"\u3010\u6a21\u578b\u6d4b\u8bd5\u3011\u5927\u6a21\u578b\u6d4b\u8bc4\u4f53\u7cfb\u7684\u6784\u6210\",\"datePublished\":\"2025-02-16T03:54:24+00:00\",\"dateModified\":\"2025-04-07T06:51:23+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/17aitech.com\\\/?p=38434\"},\"wordCount\":376,\"commentCount\":0,\"image\":{\"@id\":\"https:\\\/\\\/17aitech.com\\\/?p=38434#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/17aitech.com\\\/wp-content\\\/uploads\\\/2025\\\/02\\\/\u5c01\u9762.png\",\"keywords\":[\"\u5927\u6a21\u578b\",\"\u6a21\u578b\u6d4b\u8bd5\"],\"articleSection\":[\"\u535a\u5ba2\u6587\u7ae0\"],\"inLanguage\":\"zh-Hans\",\"potentialAction\":[{\"@type\":\"CommentAction\",\"name\":\"Comment\",\"target\":[\"https:\\\/\\\/17aitech.com\\\/?p=38434#respond\"]}]},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/17aitech.com\\\/?p=38434\",\"url\":\"https:\\\/\\\/17aitech.com\\\/?p=38434\",\"name\":\"\u3010\u6a21\u578b\u6d4b\u8bd5\u3011\u5927\u6a21\u578b\u6d4b\u8bc4\u4f53\u7cfb\u7684\u6784\u6210 - \u4e00\u8d77AI\u6280\u672f\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/17aitech.com\\\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\\\/\\\/17aitech.com\\\/?p=38434#primaryimage\"},\"image\":{\"@id\":\"https:\\\/\\\/17aitech.com\\\/?p=38434#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/17aitech.com\\\/wp-content\\\/uploads\\\/2025\\\/02\\\/\u5c01\u9762.png\",\"datePublished\":\"2025-02-16T03:54:24+00:00\",\"dateModified\":\"2025-04-07T06:51:23+00:00\",\"author\":{\"@id\":\"https:\\\/\\\/17aitech.com\\\/#\\\/schema\\\/person\\\/3d23bb6f7f115fcefc9ae7803a691739\"},\"breadcrumb\":{\"@id\":\"https:\\\/\\\/17aitech.com\\\/?p=38434#breadcrumb\"},\"inLanguage\":\"zh-Hans\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/17aitech.com\\\/?p=38434\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"zh-Hans\",\"@id\":\"https:\\\/\\\/17aitech.com\\\/?p=38434#primaryimage\",\"url\":\"https:\\\/\\\/17aitech.com\\\/wp-content\\\/uploads\\\/2025\\\/02\\\/\u5c01\u9762.png\",\"contentUrl\":\"https:\\\/\\\/17aitech.com\\\/wp-content\\\/uploads\\\/2025\\\/02\\\/\u5c01\u9762.png\",\"width\":1024,\"height\":1024},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/17aitech.com\\\/?p=38434#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"\u9996\u9875\",\"item\":\"https:\\\/\\\/17aitech.com\\\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"\u3010\u6a21\u578b\u6d4b\u8bd5\u3011\u5927\u6a21\u578b\u6d4b\u8bc4\u4f53\u7cfb\u7684\u6784\u6210\"}]},{\"@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\\\/3d23bb6f7f115fcefc9ae7803a691739\",\"name\":\"Dongming\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"zh-Hans\",\"@id\":\"\\\/\\\/17aitech.com\\\/wp-content\\\/uploads\\\/member\\\/avatars\\\/238a0b923820dcc5.1732798681.jpg\",\"url\":\"\\\/\\\/17aitech.com\\\/wp-content\\\/uploads\\\/member\\\/avatars\\\/238a0b923820dcc5.1732798681.jpg\",\"contentUrl\":\"\\\/\\\/17aitech.com\\\/wp-content\\\/uploads\\\/member\\\/avatars\\\/238a0b923820dcc5.1732798681.jpg\",\"caption\":\"Dongming\"},\"description\":\"\u89c1\u5929\u5730\uff0c\u89c1\u4f17\u751f\uff0c\u89c1\u81ea\u5df1\u3002\",\"sameAs\":[\"http:\\\/\\\/17aitech.com\"],\"url\":\"https:\\\/\\\/17aitech.com\\\/?page_id=33738&user=1\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"\u3010\u6a21\u578b\u6d4b\u8bd5\u3011\u5927\u6a21\u578b\u6d4b\u8bc4\u4f53\u7cfb\u7684\u6784\u6210 - \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=38434","schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/17aitech.com\/?p=38434#article","isPartOf":{"@id":"https:\/\/17aitech.com\/?p=38434"},"author":{"name":"Dongming","@id":"https:\/\/17aitech.com\/#\/schema\/person\/3d23bb6f7f115fcefc9ae7803a691739"},"headline":"\u3010\u6a21\u578b\u6d4b\u8bd5\u3011\u5927\u6a21\u578b\u6d4b\u8bc4\u4f53\u7cfb\u7684\u6784\u6210","datePublished":"2025-02-16T03:54:24+00:00","dateModified":"2025-04-07T06:51:23+00:00","mainEntityOfPage":{"@id":"https:\/\/17aitech.com\/?p=38434"},"wordCount":376,"commentCount":0,"image":{"@id":"https:\/\/17aitech.com\/?p=38434#primaryimage"},"thumbnailUrl":"https:\/\/17aitech.com\/wp-content\/uploads\/2025\/02\/\u5c01\u9762.png","keywords":["\u5927\u6a21\u578b","\u6a21\u578b\u6d4b\u8bd5"],"articleSection":["\u535a\u5ba2\u6587\u7ae0"],"inLanguage":"zh-Hans","potentialAction":[{"@type":"CommentAction","name":"Comment","target":["https:\/\/17aitech.com\/?p=38434#respond"]}]},{"@type":"WebPage","@id":"https:\/\/17aitech.com\/?p=38434","url":"https:\/\/17aitech.com\/?p=38434","name":"\u3010\u6a21\u578b\u6d4b\u8bd5\u3011\u5927\u6a21\u578b\u6d4b\u8bc4\u4f53\u7cfb\u7684\u6784\u6210 - \u4e00\u8d77AI\u6280\u672f","isPartOf":{"@id":"https:\/\/17aitech.com\/#website"},"primaryImageOfPage":{"@id":"https:\/\/17aitech.com\/?p=38434#primaryimage"},"image":{"@id":"https:\/\/17aitech.com\/?p=38434#primaryimage"},"thumbnailUrl":"https:\/\/17aitech.com\/wp-content\/uploads\/2025\/02\/\u5c01\u9762.png","datePublished":"2025-02-16T03:54:24+00:00","dateModified":"2025-04-07T06:51:23+00:00","author":{"@id":"https:\/\/17aitech.com\/#\/schema\/person\/3d23bb6f7f115fcefc9ae7803a691739"},"breadcrumb":{"@id":"https:\/\/17aitech.com\/?p=38434#breadcrumb"},"inLanguage":"zh-Hans","potentialAction":[{"@type":"ReadAction","target":["https:\/\/17aitech.com\/?p=38434"]}]},{"@type":"ImageObject","inLanguage":"zh-Hans","@id":"https:\/\/17aitech.com\/?p=38434#primaryimage","url":"https:\/\/17aitech.com\/wp-content\/uploads\/2025\/02\/\u5c01\u9762.png","contentUrl":"https:\/\/17aitech.com\/wp-content\/uploads\/2025\/02\/\u5c01\u9762.png","width":1024,"height":1024},{"@type":"BreadcrumbList","@id":"https:\/\/17aitech.com\/?p=38434#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"\u9996\u9875","item":"https:\/\/17aitech.com\/"},{"@type":"ListItem","position":2,"name":"\u3010\u6a21\u578b\u6d4b\u8bd5\u3011\u5927\u6a21\u578b\u6d4b\u8bc4\u4f53\u7cfb\u7684\u6784\u6210"}]},{"@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\/3d23bb6f7f115fcefc9ae7803a691739","name":"Dongming","image":{"@type":"ImageObject","inLanguage":"zh-Hans","@id":"\/\/17aitech.com\/wp-content\/uploads\/member\/avatars\/238a0b923820dcc5.1732798681.jpg","url":"\/\/17aitech.com\/wp-content\/uploads\/member\/avatars\/238a0b923820dcc5.1732798681.jpg","contentUrl":"\/\/17aitech.com\/wp-content\/uploads\/member\/avatars\/238a0b923820dcc5.1732798681.jpg","caption":"Dongming"},"description":"\u89c1\u5929\u5730\uff0c\u89c1\u4f17\u751f\uff0c\u89c1\u81ea\u5df1\u3002","sameAs":["http:\/\/17aitech.com"],"url":"https:\/\/17aitech.com\/?page_id=33738&user=1"}]}},"_links":{"self":[{"href":"https:\/\/17aitech.com\/index.php?rest_route=\/wp\/v2\/posts\/38434","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\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/17aitech.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=38434"}],"version-history":[{"count":14,"href":"https:\/\/17aitech.com\/index.php?rest_route=\/wp\/v2\/posts\/38434\/revisions"}],"predecessor-version":[{"id":39958,"href":"https:\/\/17aitech.com\/index.php?rest_route=\/wp\/v2\/posts\/38434\/revisions\/39958"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/17aitech.com\/index.php?rest_route=\/wp\/v2\/media\/38435"}],"wp:attachment":[{"href":"https:\/\/17aitech.com\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=38434"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/17aitech.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=38434"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/17aitech.com\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=38434"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}