{"id":2433,"date":"2023-07-26T09:59:52","date_gmt":"2023-07-26T01:59:52","guid":{"rendered":""},"modified":"2023-07-26T09:59:52","modified_gmt":"2023-07-26T01:59:52","slug":"","status":"publish","type":"post","link":"http:\/\/www.idea2003.com\/dongtai\/2433.html","title":{"rendered":"\u8c01\u5728\u7ed9\u5927\u6a21\u578b\u6253\u5206\uff1f"},"content":{"rendered":"

\u5f53\u65b0\u8d5b\u9053\u6324\u6ee1\u4e86\u6469\u62f3\u64e6\u638c\u7684\u9009\u624b\uff0c\u573a\u8fb9\u7684\u88c1\u5224\u5458\u4e5f\u5e94\u8fd0\u800c\u751f\u3002<\/p>\n

5\u6708\uff0c\u56fd\u5bb6\u79d1\u6280\u90e8\u4e0b\u5c5e\u7684\u4e2d\u56fd\u79d1\u5b66\u6280\u672f\u4fe1\u606f\u7814\u7a76\u6240\uff0c\u53d1\u5e03\u4e86\u300a\u4e2d\u56fd\u4eba\u5de5\u667a\u80fd\u5927\u6a21\u578b\u5730\u56fe\u7814\u7a76\u62a5\u544a\u300b\u3002\u5185\u5bb9\u663e\u793a\uff0c\u622a\u81f35\u670828\u65e5\uff0c\u56fd\u518510\u4ebf\u7ea7\u53c2\u6570\u89c4\u6a21\u4ee5\u4e0a\u57fa\u7840\u5927\u6a21\u578b\u81f3\u5c11\u5df2\u53d1\u5e0379\u4e2a\u3002<\/p>\n

\u6bcf\u4e00\u4e2a\u4eae\u76f8\u65f6\uff0c\u90fd\u5c11\u4e0d\u4e86\u201c\u884c\u4e1a\u9886\u5148\u201d\u201c\u6280\u672f\u9769\u65b0\u201d\u8bf8\u5982\u6b64\u7c7b\u7684\u6807\u7b7e\u3002\u4e0d\u514d\u5f15\u6765\u8d28\u7591\uff1a\u5982\u4f55\u76f4\u89c2\u5730\u8bc4\u5224\u54ea\u4e00\u6b3e\u5927\u6a21\u578b\u5728\u6280\u672f\u548c\u6027\u80fd\u4e0a\u66f4\u4e3a\u5353\u8d8a\uff1f\u90a3\u4e9b\u5ba3\u79f0\u201c\u7b2c\u4e00\u201d\u7684\u8bc4\u4f30\u6807\u51c6\u4e0e\u6570\u636e\u6765\u6e90\u53c8\u662f\u600e\u6837\u7684\uff1f<\/p>\n

\u4e00\u628a\u8861\u91cf\u4e0d\u540c\u6a21\u578b\u6548\u80fd\u57fa\u51c6\u7684\u201c\u5c3a\u5b50\u201d\u4e9f\u5f85\u6253\u9020\u3002<\/p>\n

\u524d\u4e0d\u4e45\uff0c\u56fd\u9645\u54a8\u8be2\u516c\u53f8IDC\u53d1\u5e03\u300aAI\u5927\u6a21\u578b\u6280\u672f\u80fd\u529b\u8bc4\u4f30\u62a5\u544a2023\u300b\uff0c\u8c03\u7814\u4e869\u5bb6\u4e2d\u56fd\u5e02\u573a\u4e3b\u6d41\u5927\u6a21\u578b\u6280\u672f\u5382\u5546\u3002\u5176\u4ed6\u4e0d\u5c11\u7814\u7a76\u673a\u6784\u548c\u56e2\u961f\u4e5f\u6295\u5165\u8d44\u6e90\uff0c\u53d1\u5e03\u4e86\u5bf9\u5e94\u7684\u8bc4\u4ef7\u6807\u51c6\u548c\u6df1\u5ea6\u62a5\u544a\u3002\u8fd9\u80cc\u540e\u6240\u663e\u9732\u7684\u73b0\u8c61\u548c\u8d8b\u52bf\uff0c\u66f4\u503c\u5f97\u6df1\u5c42\u6b21\u7684\u63a2\u8ba8\u3002<\/p>\n

\u8bc4\u6d4b\u57fa\u51c6\u767e\u5bb6\u4e89\u9e23<\/strong><\/h2>\n

ChatGPT \u5e26\u706b\u4e86\u5927\u6a21\u578b\u5e94\u7528\u7684\u76f8\u5173\u7814\u7a76\uff0c\u8bc4\u6d4b\u57fa\u51c6\u4ea6\u6210\u4e3a\u5173\u6ce8\u7126\u70b9\u6240\u5728\u3002<\/p>\n

\u65e5\u524d\uff0c\u5fae\u8f6f\u4e9a\u6d32\u7814\u7a76\u9662\u516c\u5f00\u4e86\u4ecb\u7ecd\u5927\u6a21\u578b\u8bc4\u6d4b\u9886\u57df\u7684\u7efc\u8ff0\u6587\u7ae0\u300aA Survey on Evaluation of Large Language Models\u300b\u3002\u6839\u636e\u4e0d\u5b8c\u5168\u7edf\u8ba1\uff08\u89c1\u4e0b\u56fe\uff09\uff0c\u5927\u6a21\u578b\u8bc4\u6d4b\u65b9\u9762\u6587\u7ae0\u7684\u53d1\u8868\u5448\u4e0a\u5347\u8d8b\u52bf\uff0c\u8d8a\u6765\u8d8a\u591a\u7684\u7814\u7a76\u7740\u773c\u4e8e\u8bbe\u8ba1\u66f4\u79d1\u5b66\u3001\u66f4\u597d\u5ea6\u91cf\u3001\u66f4\u51c6\u786e\u7684\u8bc4\u6d4b\u65b9\u5f0f\u6765\u5bf9\u5927\u6a21\u578b\u7684\u80fd\u529b\u8fdb\u884c\u66f4\u6df1\u5165\u7684\u4e86\u89e3\u3002<\/p>\n

\u6587\u4e2d\u4e00\u5171\u5217\u51fa\u4e8619\u4e2a\u53d7\u6b22\u8fce\u7684\u57fa\u51c6\u6d4b\u8bd5\uff0c\u6bcf\u4e2a\u90fd\u4fa7\u91cd\u4e8e\u4e0d\u540c\u7684\u65b9\u9762\u548c\u8bc4\u4f30\u6807\u51c6\uff0c\u4e3a\u5176\u5404\u81ea\u7684\u9886\u57df\u63d0\u4f9b\u4e86\u5b9d\u8d35\u7684\u8d21\u732e\u3002\u4e3a\u4e86\u66f4\u597d\u5730\u603b\u7ed3\uff0c\u7814\u7a76\u5458\u5c06\u8fd9\u4e9b\u57fa\u51c6\u6d4b\u8bd5\u5206\u4e3a\u4e24\u7c7b\uff1a\u901a\u7528\u57fa\u51c6\uff08General benchmarks\uff09\u548c\u5177\u4f53\u57fa\u51c6\uff08Specific benchmarks\uff09\uff0c\u5176\u4e2d\u4e0d\u4e4f\u4e00\u4e9b\u6df1\u5177\u76db\u540d\u7684\u5927\u6a21\u578b\u57fa\u51c6\u3002<\/p>\n

Chatbot Arena\uff0c\u5c31\u88ab\u884c\u4e1a\u4eba\u58eb\u666e\u904d\u8ba4\u4e3a\u662f\u6700\u5177\u516c\u5e73\u6027\u4e0e\u5e7f\u6cdb\u63a5\u53d7\u5ea6\u7684\u5e73\u53f0\u3002\u5176\u80cc\u540e\u7684\u63a8\u624b\u2014\u2014LMSYS Org\uff0c\u662f\u4e00\u4e2a\u5f00\u653e\u7684\u7814\u7a76\u7ec4\u7ec7\uff0c\u7531\u52a0\u5dde\u5927\u5b66\u4f2f\u514b\u5229\u5206\u6821\u3001\u52a0\u5dde\u5927\u5b66\u5723\u5730\u4e9a\u54e5\u5206\u6821\u548c\u5361\u5185\u57fa\u6885\u9686\u5927\u5b66\u5408\u4f5c\u521b\u7acb\u3002<\/p>\n

\u8fd9\u4e2a\u521b\u65b0\u6027\u7684\u8bc4\u4f30\u6807\u51c6\uff0c\u4e3a\u5404\u5927AI\u7814\u7a76\u673a\u6784\u4e0e\u6280\u672f\u7231\u597d\u8005\uff0c\u63d0\u4f9b\u4e86\u4e00\u4e2a\u65e2\u72ec\u7279\u53c8\u5177\u6709\u6fc0\u70c8\u7ade\u4e89\u529b\u7684\u573a\u6240\uff0c\u4e13\u95e8\u7528\u4e8e\u8bc4\u4ef7\u548c\u6bd4\u5bf9\u4e0d\u540c\u804a\u5929\u673a\u5668\u4eba\u6a21\u578b\u7684\u5b9e\u9645\u5e94\u7528\u6548\u679c\u3002\u7528\u6237\u80fd\u591f\u4e0e\u5176\u4e2d\u7684\u533f\u540d\u6a21\u578b\u8fdb\u884c\u5b9e\u65f6\u4e92\u52a8\uff0c\u800c\u540e\u901a\u8fc7\u5728\u7ebf\u6295\u7968\u7cfb\u7edf\u8868\u8fbe\u4ed6\u4eec\u5bf9\u4e8e\u67d0\u4e00\u6a21\u578b\u7684\u6ee1\u610f\u5ea6\u6216\u559c\u597d\u3002<\/p>\n

\u503c\u5f97\u4e00\u63d0\u7684\u662f\uff0c\u8be5\u8bc4\u6d4b\u65b9\u5f0f\u7684\u8bbe\u8ba1\u7075\u611f\u6765\u6e90\u4e8e\u56fd\u9645\u8c61\u68cb\u7b49\u7ade\u6280\u6e38\u620f\u4e2d\u76db\u884c\u7684ElO\u8bc4\u5206\u7cfb\u7edf\u3002\u901a\u8fc7\u79ef\u7d2f\u5927\u91cf\u7684\u7528\u6237\u6295\u7968\uff0c\u5b83\u80fd\u591f\u66f4\u4e3a\u8d34\u8fd1\u5b9e\u9645\u573a\u666f\u5730\u8bc4\u4f30\u5404\u6a21\u578b\u7684\u7efc\u5408\u8868\u73b0\u3002<\/p>\n

\u5fae\u8f6f\u4e9a\u6d32\u7814\u7a76\u9662\u8fd8\u5728\u6587\u4e2d\u63d0\u5230\u4e86\u901a\u7528\u57fa\u51c6C-Eval\uff0c\u8fd9\u662f\u4e00\u4e2a\u5168\u9762\u7684\u4e2d\u6587\u57fa\u7840\u6a21\u578b\u8bc4\u4f30\u5957\u4ef6\u3002\u5b83\u5305\u542b\u4e8613948\u4e2a\u591a\u9879\u9009\u62e9\u9898\uff0c\u6db5\u76d6\u4e8652\u4e2a\u4e0d\u540c\u7684\u5b66\u79d1\u548c\u56db\u4e2a\u96be\u5ea6\u7ea7\u522b\uff0c\u8be5\u9879\u76ee\u7531\u4e0a\u6d77\u4ea4\u901a\u5927\u5b66\u3001\u6e05\u534e\u5927\u5b66\u3001\u7231\u4e01\u5821\u5927\u5b66\u5171\u540c\u5b8c\u6210\u3002<\/p>\n

\u9664\u4e86\u901a\u7528\u4efb\u52a1\u7684\u57fa\u51c6\u6d4b\u8bd5\u5916\uff0c\u8fd8\u5b58\u5728\u4e00\u4e9b\u4e13\u4e3a\u67d0\u4e9b\u4e0b\u6e38\u4efb\u52a1\u8bbe\u8ba1\u7684\u5177\u4f53\u57fa\u51c6\u6d4b\u8bd5\u3002<\/p>\n

\u8b6c\u5982\uff0cMultiMedQA\u662f\u4e00\u4e2a\u533b\u5b66\u95ee\u7b54\u57fa\u51c6\u6d4b\u8bd5\uff0c\u91cd\u70b9\u5173\u6ce8\u533b\u5b66\u68c0\u67e5\u3001\u533b\u5b66\u7814\u7a76\u548c\u6d88\u8d39\u8005\u5065\u5eb7\u95ee\u9898\u3002\u8be5\u57fa\u51c6\u7531\u8c37\u6b4c\u548cDeepMind\u7684\u79d1\u7814\u4eba\u5458\u63d0\u51fa\uff0c\u5b83\u5305\u62ec\u4e03\u4e2a\u4e0e\u533b\u5b66\u95ee\u7b54\u76f8\u5173\u7684\u6570\u636e\u96c6\uff0c\u5176\u4e2d\u5305\u62ec\u516d\u4e2a\u73b0\u6709\u7684\u6570\u636e\u96c6\u548c\u4e00\u4e2a\u65b0\u7684\u6570\u636e\u96c6\u3002\u6d4b\u8bd5\u76ee\u6807\u662f\u8bc4\u4f30\u5927\u8bed\u8a00\u6a21\u578b\u5728\u4e34\u5e8a\u77e5\u8bc6\u548c\u95ee\u7b54\u80fd\u529b\u65b9\u9762\u7684\u6027\u80fd\u3002<\/p>\n

\u8fd8\u6709\u4e00\u4e9b\u4e2d\u6587\u8bc4\u6d4b\u57fa\u51c6\u88ab\u5fae\u8f6f\u7814\u7a76\u9662\u6240\u9057\u6f0f\u3002\u4f8b\u5982SuperCLUE\uff0c\u4f5c\u4e3a\u9488\u5bf9\u4e2d\u6587\u53ef\u7528\u7684\u901a\u7528\u5927\u6a21\u578b\u7684\u4e00\u4e2a\u6d4b\u8bc4\u57fa\u51c6\uff0c\u7531\u6765\u81ea\u4e2d\u6587\u8bed\u8a00\u7406\u89e3\u6d4b\u8bc4\u57fa\u51c6\u5f00\u6e90\u793e\u533aCLUE\u7684\u6210\u5458\u53d1\u8d77\u3002<\/p>\n

\u4e3a\u4e86\u7740\u773c\u4e8e\u7efc\u5408\u8bc4\u4ef7\u5927\u6a21\u578b\u7684\u80fd\u529b\uff0c\u4f7f\u5176\u80fd\u5168\u9762\u5730\u6d4b\u8bd5\u5927\u6a21\u578b\u7684\u6548\u679c\uff0c\u53c8\u80fd\u8003\u5bdf\u6a21\u578b\u5728\u4e2d\u6587\u4e0a\u7279\u6709\u4efb\u52a1\u7684\u7406\u89e3\u548c\u79ef\u7d2f\uff0cSuperCLUE\u4ece\u4e09\u4e2a\u4e0d\u540c\u7684\u7ef4\u5ea6\u8bc4\u4ef7\u6a21\u578b\u7684\u80fd\u529b\uff1a\u57fa\u7840\u80fd\u529b\u3001\u4e13\u4e1a\u80fd\u529b\u548c\u4e2d\u6587\u7279\u6027\u80fd\u529b\u3002<\/p>\n

\u5b9e\u9645\u4e0a\uff0c\u8fd9\u4e9b\u7ef4\u5ea6\u53ea\u662f\u51b0\u5c71\u4e00\u89d2\u3002\u5728\u8bc4\u6d4b LLMs \u7684\u6027\u80fd\u65f6\uff0c\u9009\u62e9\u5408\u9002\u7684\u4efb\u52a1\u548c\u9886\u57df\u5bf9\u4e8e\u5c55\u793a\u5927\u578b\u8bed\u8a00\u6a21\u578b\u7684\u8868\u73b0\u3001\u4f18\u52bf\u548c\u52a3\u52bf\u81f3\u5173\u91cd\u8981\u3002\u5fae\u8f6f\u4e9a\u6d32\u7814\u7a76\u9662\u5c06\u73b0\u6709\u7684\u4efb\u52a1\u5212\u5206\u4e3a\u4e867\u4e2a\u4e0d\u540c\u7684\u7c7b\u522b\uff1a<\/p>\n

\u81ea\u7136\u8bed\u8a00\u5904\u7406\uff1a\u5305\u62ec\u81ea\u7136\u8bed\u8a00\u7406\u89e3\u3001\u63a8\u7406\u3001\u81ea\u7136\u8bed\u8a00\u751f\u6210\u548c\u591a\u8bed\u8a00\u4efb\u52a1\uff1b<\/p>\n

\u9c81\u68d2\u6027\u3001\u4f26\u7406\u3001\u504f\u89c1\u548c\u771f\u5b9e\u6027\uff1b<\/p>\n

\u533b\u5b66\u5e94\u7528\uff1a\u5305\u62ec\u533b\u5b66\u95ee\u7b54\u3001\u533b\u5b66\u8003\u8bd5\u3001\u533b\u5b66\u6559\u80b2\u548c\u533b\u5b66\u52a9\u624b\uff1b<\/p>\n

\u793e\u4f1a\u79d1\u5b66\uff1b<\/p>\n

\u81ea\u7136\u79d1\u5b66\u4e0e\u5de5\u7a0b\uff1a\u5305\u62ec\u6570\u5b66\u3001\u901a\u7528\u79d1\u5b66\u548c\u5de5\u7a0b\uff1b<\/p>\n

\u4ee3\u7406\u5e94\u7528\uff1a\u5c06 LLMs \u4f5c\u4e3a\u4ee3\u7406\u4f7f\u7528\uff1b<\/p>\n

\u5176\u4ed6\u5e94\u7528\u3002<\/p>\n

\u7eb5\u89c2\u5f53\u524dAI\u9886\u57df\u7684\u53d1\u5c55\u8d8b\u52bf\uff0c\u5927\u6a21\u578b\u7684\u8bc4\u6d4b\u57fa\u51c6\u6d4b\u8bd5\u4e0d\u518d\u4ec5\u4ec5\u662f\u4e00\u4e2a\u5355\u4e00\u7684\u6280\u672f\u73af\u8282\uff0c\u800c\u662f\u5df2\u7ecf\u9010\u6b65\u6210\u4e3a\u6574\u4e2a\u4e0a\u4e0b\u6e38\u4ea7\u4e1a\u94fe\u4e2d\u7684\u91cd\u8981\u914d\u5957\u3002<\/p>\n

\u6b63\u786e\u5730\u4f7f\u7528\u201c\u5c3a\u5b50\u201d<\/strong><\/h2>\n

\u5b58\u5728\u5373\u5408\u7406\u3002<\/p>\n

\u5927\u6a21\u578b\u8bc4\u6d4b\u57fa\u51c6\u7684\u8bde\u751f\u548c\u6301\u7eed\u4f18\u5316\uff0c\u4e4b\u6240\u4ee5\u5448\u73b0\u51fa\u6108\u6f14\u6108\u70c8\u7684\u52bf\u5934\uff0c\u65e0\u7591\u4e0e\u5176\u6240\u80fd\u5e26\u6765\u7684\u5de8\u5927\u4ef7\u503c\u548c\u4e1a\u754c\u7684\u5e7f\u6cdb\u8ba4\u540c\u662f\u5bc6\u4e0d\u53ef\u5206\u7684\u3002<\/p>\n

\u53ef\u4ee5\u770b\u5230\uff0c\u901a\u8fc7\u6df1\u5165\u7684\u5927\u6a21\u578b\u8bc4\u6d4b\u57fa\u51c6\u5206\u6790\uff0c\u80fd\u591f\u66f4\u4e3a\u660e\u786e\u548c\u7cfb\u7edf\u5730\u63ed\u793a\u5927\u6a21\u578b\u5728\u5404\u79cd\u5e94\u7528\u573a\u666f\u4e2d\u7684\u4f18\u52bf\u4e0e\u5c40\u9650\u6027\u3002\u8fd9\u79cd\u4e13\u4e1a\u7684\u8bc4\u4f30\u4e0d\u4ec5\u4e3aAI\u9886\u57df\u7684\u7814\u53d1\u8005\u63d0\u4f9b\u4e86\u6e05\u6670\u7684\u6307\u5bfc\uff0c\u540c\u65f6\u4e5f\u52a9\u529b\u7528\u6237\u6700\u7ec8\u4f5c\u51fa\u66f4\u4e3a\u660e\u667a\u7684\u6280\u672f\u9009\u62e9\u3002<\/p>\n

\u5728\u590d\u6742\u7684\u7814\u53d1\u8fc7\u7a0b\u4e2d\uff0c\u5224\u65ad\u6280\u672f\u65b9\u6848\u6216\u7279\u5b9a\u6a21\u578b\u7684\u4f18\u8d8a\u6027\u5f80\u5f80\u662f\u4e00\u4e2a\u6311\u6218\u3002C-Eval\u6570\u636e\u96c6\u548c\u5176\u76f8\u5173\u699c\u5355\uff0c\u610f\u4e49\u4e0d\u4ec5\u4ec5\u662f\u4e00\u7cfb\u5217\u7684\u6570\u5b57\u6216\u6392\u540d\uff0c\u800c\u662f\u4e3a\u5927\u6a21\u578b\u7684\u7814\u53d1\u8005\u63d0\u4f9b\u4e86\u4e00\u5957\u5ba2\u89c2\u3001\u7cfb\u7edf\u7684\u8bc4\u4f30\u5de5\u5177\u3002<\/p>\n

\u7528C-Eval\u9879\u76ee\u56e2\u961f\u7684\u8bdd\u6765\u8bf4\uff0c\u201c\u6211\u4eec\u7684\u6700\u91cd\u8981\u76ee\u6807\u662f\u8f85\u52a9\u6a21\u578b\u5f00\u53d1\u201d\u3002<\/p>\n

\u5177\u4f53\u6765\u770b\uff0c\u7814\u53d1\u56e2\u961f\u53ef\u4ee5\u4e0e\u4f01\u4e1a\u7d27\u5bc6\u5408\u4f5c\uff0c\u5c06\u5927\u6a21\u578b\u8bc4\u6d4b\u57fa\u51c6\u6574\u5408\u5230\u4ed6\u4eec\u7684\u5f00\u53d1\u548c\u6d4b\u8bd5\u5de5\u4f5c\u6d41\u7a0b\u4e2d\u3002\u8fd9\u4e0d\u4ec5\u53ef\u4ee5\u5728\u5b9e\u9645\u5e94\u7528\u73af\u5883\u4e2d\u9a8c\u8bc1\u6a21\u578b\u7684\u6027\u80fd\uff0c\u8fd8\u80fd\u901a\u8fc7\u53cc\u65b9\u7684\u6df1\u5ea6\u6c9f\u901a\uff0c\u627e\u5230\u5728\u6d4b\u8bd5\u8fc7\u7a0b\u4e2d\u53ef\u80fd\u9047\u5230\u7684\u6280\u672f\u96be\u9898\u548c\u6311\u6218\uff0c\u4ece\u800c\u5b9e\u73b0\u66f4\u4e3a\u9ad8\u6548\u548c\u51c6\u786e\u7684\u6a21\u578b\u4f18\u5316\u3002<\/p>\n

\u6b63\u662f\u57fa\u4e8e\u8fd9\u4e00\u70b9\uff0c\u591a\u5bb6\u5934\u90e8\u5927\u6a21\u578b\u5382\u5546\u4e0d\u4ec5\u5728\u6a21\u578b\u7814\u53d1\u4e0a\u6301\u7eed\u6295\u5165\uff0c\u540c\u65f6\u4e5f\u5728\u8bc4\u6d4b\u57fa\u51c6\u7684\u5236\u5b9a\u4e0e\u4f18\u5316\u4e0a\u4e0b\u8db3\u4e86\u529f\u592b\u3002<\/p>\n

\u8b6c\u5982\u79d1\u5927\u8baf\u98de\u901a\u8fc7\u8ba4\u77e5\u667a\u80fd\u5168\u56fd\u91cd\u70b9\u5b9e\u9a8c\u5ba4\u7275\u5934\u8bbe\u8ba1\u4e86\u901a\u7528\u8ba4\u77e5\u5927\u6a21\u578b\u8bc4\u6d4b\u4f53\u7cfb\uff0c\u8986\u76d67\u5927\u7c7b481\u4e2a\u7ec6\u5206\u4efb\u52a1\u7c7b\u578b\uff1b\u963f\u91cc\u5df4\u5df4\u8fbe\u6469\u9662\u591a\u8bed\u8a00NLP\u56e2\u961f\u53d1\u5e03\u4e86\u9996\u4e2a\u591a\u8bed\u8a00\u591a\u6a21\u6001\u6d4b\u8bd5\u57fa\u51c6M3Exam\uff0c\u5171\u6db5\u76d612317\u9053\u9898\u76ee\uff0c\u7b49\u7b49\u3002<\/p>\n

\u4e0d\u8fc7\u4e5f\u6b63\u5982C-Eval\u9879\u76ee\u56e2\u961f\u6240\u5f3a\u8c03\u7684\uff1a\u5bf9\u4e8e\u5927\u6a21\u578b\u5382\u5546\uff0c\u5355\u7eaf\u5730\u8ffd\u6c42\u699c\u5355\u7684\u9ad8\u4f4d\u6392\u540d\uff0c\u5e76\u4e0d\u5e94\u6210\u4e3a\u5176\u4e3b\u8981\u8ffd\u6c42\u3002<\/p>\n

\u5f53\u5382\u5546\u5c06\u699c\u5355\u6210\u7ee9\u4f5c\u4e3a\u9996\u8981\u76ee\u6807\u65f6\uff0c\u53ef\u80fd\u4f1a\u4e3a\u4e86\u8ffd\u6c42\u9ad8\u5206\u800c\u91c7\u7528\u8fc7\u5ea6\u62df\u5408\u699c\u5355\u7684\u7b56\u7565\uff0c\u8fd9\u6837\u5c31\u5f88\u5bb9\u6613\u635f\u5931\u6a21\u578b\u7684\u5e7f\u6cdb\u9002\u7528\u6027\u3002\u66f4\u4e3a\u5173\u952e\u7684\u662f\uff0c\u82e5\u4ec5\u4ec5\u7740\u773c\u4e8e\u6392\u540d\uff0c\u5382\u5546\u53ef\u80fd\u4e3a\u4e86\u77ed\u671f\u7684\u6210\u7ee9\u800c\u8bd5\u56fe\u5bfb\u627e\u6377\u5f84\uff0c\u4ece\u800c\u8fdd\u80cc\u4e86\u771f\u6b63\u8e0f\u5b9e\u7684\u79d1\u7814\u7cbe\u795e\u4e0e\u4ef7\u503c\u89c2\u3002<\/p>\n

\u518d\u770b\u7ec8\u7aef\u7528\u6237\u7684\u89c6\u89d2\u4e2d\uff0c\u5927\u6a21\u578b\u6d4b\u8bc4\u57fa\u51c6\u63d0\u4f9b\u4e86\u4e00\u4e2a\u5168\u9762\u7684\u3001\u7ed3\u6784\u5316\u7684\u53c2\u8003\u6846\u67b6\uff0c\u4ece\u800c\u5145\u5206\u5730\u8f85\u52a9\u7528\u6237\u5728\u4f17\u591a\u6280\u672f\u9009\u9879\u4e2d\u505a\u51fa\u66f4\u4e3a\u7406\u6027\u548c\u660e\u667a\u7684\u51b3\u7b56\u3002\u8fd9\u79cd\u8bc4\u6d4b\u4e0d\u4ec5\u964d\u4f4e\u4e86\u6280\u672f\u91c7\u7eb3\u7684\u98ce\u9669\uff0c\u4e5f\u786e\u4fdd\u4e86\u7528\u6237\u80fd\u591f\u4ece\u6240\u9009\u6a21\u578b\u4e2d\u83b7\u5f97\u6700\u4f73\u7684\u6295\u8d44\u56de\u62a5\u7387\u3002<\/p>\n

\u5c24\u5176\u5bf9\u4e8e\u90a3\u4e9b\u8fd8\u672a\u62e5\u6709\u6df1\u539a\u5927\u6a21\u578b\u7814\u53d1\u5b9e\u529b\u7684\u4f01\u4e1a\u6765\u8bf4\uff0c\u6df1\u5165\u4e86\u89e3\u5927\u6a21\u578b\u7684\u6280\u672f\u8fb9\u754c\uff0c\u5e76\u80fd\u591f\u9488\u5bf9\u81ea\u8eab\u9700\u6c42\u9ad8\u6548\u5730\u8fdb\u884c\u6280\u672f\u9009\u578b\uff0c\u662f\u81f3\u5173\u91cd\u8981\u7684\u3002<\/p>\n

\u7efc\u4e0a\uff0c\u4e0d\u8bba\u662f\u5bf9\u4e8e\u80cc\u540e\u7684\u7814\u53d1\u56e2\u961f\u8fd8\u662f\u4ea7\u54c1\u4fa7\u7684\u7ec8\u7aef\u7528\u6237\uff0c\u5927\u6a21\u578b\u8bc4\u6d4b\u57fa\u51c6\u90fd\u627f\u8f7d\u7740\u4e0d\u53ef\u4f30\u91cf\u7684\u4ef7\u503c\u548c\u610f\u4e49\u3002<\/p>\n

\u52a3\u5e01\u6765\u88ad<\/strong><\/h2>\n

\u540a\u8be1\u7684\u662f\uff0c\u7531\u4e8e\u5728\u539f\u7406\u6838\u5fc3\u4e0a\u5e76\u4e0d\u6d89\u53ca\u590d\u6742\u7684\u6280\u672f\u95e8\u69db\uff0c\u5bfc\u81f4\u76ee\u524d\u5e02\u573a\u4e0a\u7684\u5927\u6a21\u578b\u8bc4\u6d4b\u57fa\u51c6\u7684\u6570\u91cf\uff0c\u751a\u81f3\u5df2\u7ecf\u8d85\u8fc7\u4e86\u5927\u6a21\u578b\u672c\u8eab\u3002\u8fd9\u5176\u4e2d\u81ea\u7136\u6709\u8bb8\u591a\u673a\u6784\u89c1\u5230\u4e86\u53ef\u4e58\u4e4b\u673a\uff0c\u8fdb\u884c\u5404\u79cd\u5e02\u573a\u64cd\u4f5c\uff0c\u5305\u62ec\u6df7\u6dc6\u89c6\u542c\u3001\u8bef\u5bfc\u6d88\u8d39\u8005\u7684\u884c\u4e3a\u3002<\/p>\n

\u6b64\u524d\u5c31\u6709\u89c2\u70b9\u8ba4\u4e3a\uff0c\u968f\u7740AI\u6280\u672f\u7684\u53d1\u5c55\uff0c\u5927\u6a21\u578b\u8bc4\u6d4b\u53ef\u80fd\u4f1a\u88ab\u67d0\u4e9b\u516c\u53f8\u6216\u673a\u6784\u7528\u4f5c\u8425\u9500\u5de5\u5177\uff0c\u901a\u8fc7\u53d1\u5e03\u5176\u6a21\u578b\u7684\u9ad8\u5206\u8bc4\u6d4b\u7ed3\u679c\u6765\u5438\u5f15\u516c\u4f17\u7684\u6ce8\u610f\u529b\uff0c\u4ee5\u671f\u63d0\u9ad8\u4ea7\u54c1\u7684\u5e02\u573a\u4efd\u989d\u3002<\/p>\n

\u76ee\u524d\u4e5f\u6709\u4e00\u4e9b\u7a81\u51fa\u7684\u73b0\u8c61\u4f50\u8bc1\uff1a\u5728\u67d0\u4e9b\u7279\u5b9a\u8bc4\u6d4b\u699c\u5355\u4e2d\u9886\u5148\u7684\u5382\u5546\uff0c\u653e\u5230\u5176\u4ed6\u4e0d\u540c\u7684\u699c\u5355\u8bc4\u6d4b\u4e2d\uff0c\u5374\u672a\u80fd\u591f\u7ef4\u6301\u5176\u4f18\u52bf\u5730\u4f4d\u3002<\/p>\n

\u4e0d\u80fd\u6392\u9664\u5b58\u5728\u7740\u5ba2\u89c2\u539f\u56e0\u3002\u5f53\u524d\u9636\u6bb5\uff0c\u5bf9\u4e8e\u5927\u6a21\u578b\u7684\u8bc4\u4f30\u673a\u5236\u548c\u5177\u4f53\u8bc4\u6d4b\u6307\u6807\uff0c\u5e76\u6ca1\u6709\u8fbe\u5230\u4e00\u4e2a\u884c\u4e1a\u5171\u8bc6\uff0c\u66f4\u9051\u8bba\u51fa\u73b0\u7edf\u4e00\u7684\u7684\u8bc4\u6d4b\u6807\u51c6\u3002\u4e0d\u540c\u7684\u5e94\u7528\u73af\u5883\u548c\u4efb\u52a1\u6807\u51c6\uff0c\u5c31\u4f1a\u4ea7\u751f\u622a\u7136\u4e0d\u540c\u7684\u8bc4\u4ef7\u6846\u67b6\u548c\u9700\u6c42\u3002<\/p>\n

\u6b64\u5916\uff0c\u5927\u6a21\u578b\u8bc4\u6d4b\u901a\u5e38\u4f9d\u8d56\u4e8e\u4e24\u5927\u4e3b\u8981\u65b9\u6cd5\uff1a\u81ea\u52a8\u8bc4\u6d4b\u548c\u4eba\u5de5\u8bc4\u6d4b\u3002\u81ea\u52a8\u8bc4\u6d4b\u662f\u57fa\u4e8e\u7cbe\u786e\u7684\u8ba1\u7b97\u673a\u7b97\u6cd5\u548c\u4e00\u7cfb\u5217\u9884\u5b9a\u4e49\u7684\u6307\u6807\u8fdb\u884c\uff0c\u800c\u4eba\u5de5\u8bc4\u6d4b\u66f4\u591a\u5f3a\u8c03\u4eba\u7c7b\u4e13\u5bb6\u7684\u4e3b\u89c2\u89c1\u89e3\u3001\u7ecf\u9a8c\u548c\u8d28\u91cf\u5224\u5b9a\u3002<\/p>\n

\u9047\u5230\u5927\u6a21\u578b\u751f\u6210\u8bd7\u6b4c\u6216\u77ed\u6587\u8fd9\u7c7b\u4efb\u52a1\u65f6\uff0c\u4eba\u5de5\u8bc4\u6d4b\u7684\u4e3b\u89c2\u6027\u53d8\u5f97\u5c24\u4e3a\u663e\u8457\u3002\u81ea\u53e4\u201c\u6587\u65e0\u7b2c\u4e00\uff0c\u6b66\u65e0\u7b2c\u4e8c\u201d\uff0c\u4e0d\u540c\u7684\u8bc4\u5ba1\u8005\u53ef\u80fd\u4f1a\u5bf9\u540c\u4e00\u4f5c\u54c1\u7ed9\u51fa\u4e0d\u540c\u7684\u8bc4\u4ef7\u3002<\/p>\n

\u7136\u800c\uff0c\u4ece\u76f8\u5173\u641c\u7d22\u7ed3\u679c\u4e2d\u4e0d\u96be\u53d1\u73b0\uff0c\u5927\u6a21\u578b\u8bc4\u6d4b\u65e9\u5df2\u88ab\u67d0\u4e9b\u5382\u5546\u89c6\u4e3a\u4e00\u4e2a\u8425\u9500\u7684\u7ade\u6280\u573a\u3002\u6bd5\u7adf\u5728\u4e00\u4e2a\u7ade\u4e89\u6fc0\u70c8\u7684\u5e02\u573a\u4e2d\uff0c\u6bcf\u4e00\u4e2a\u5382\u5546\u90fd\u5e0c\u671b\u81ea\u5df1\u7684\u4ea7\u54c1\u80fd\u591f\u8131\u9896\u800c\u51fa\u3002<\/p>\n

\u56e0\u6b64\u6709\u5145\u5206\u7684\u52a8\u673a\u53bb\u9009\u62e9\u90a3\u4e9b\u80fd\u591f\u7a81\u663e\u81ea\u5df1\u4ea7\u54c1\u4f18\u52bf\u7684\u8bc4\u6d4b\u6307\u6807\uff0c\u800c\u5ffd\u7565\u90a3\u4e9b\u53ef\u80fd\u66b4\u9732\u5f31\u9879\u7684\u6307\u6807\u3002\u8fd9\u79cd\u9009\u62e9\u6027\u7684\u5c55\u793a\uff0c\u5373\u4f7f\u6709\u673a\u4f1a\u5e26\u6765\u77ed\u671f\u7684\u5e02\u573a\u4f18\u52bf\uff0c\u4f46\u662f\u5bf9\u4e8e\u6d88\u8d39\u8005\u548c\u6574\u4e2a\u884c\u4e1a\u6765\u8bf4\uff0c\u5fc5\u7136\u662f\u6709\u5bb3\u7684\u3002<\/p>\n

\u4e00\u65f6\u7684\u8bef\u5bfc\u4e00\u65e6\u626d\u66f2\u4e86\u5e02\u573a\u7684\u7ade\u4e89\u683c\u5c40\uff0c\u53ef\u80fd\u4f7f\u5f97\u771f\u6b63\u6709\u4ef7\u503c\u7684\u521b\u65b0\u88ab\u57cb\u6ca1\u3002\u52a3\u5e01\u9a71\u9010\u826f\u5e01\u4e4b\u4e0b\uff0c\u90a3\u4e9b\u53ea\u662f\u4e3a\u4e86\u5ba3\u4f20\u800c\u8fdb\u884c\u7684\u201c\u521b\u65b0\u201d\u53cd\u800c\u4f1a\u8d8b\u4e4b\u82e5\u9e5c\u3002<\/p>\n

\u4ece\u8fd9\u4e2a\u89d2\u5ea6\u51fa\u53d1\uff0c\u5927\u6a21\u578b\u8bc4\u6d4b\u57fa\u51c6\u8fd8\u662f\u5e94\u8be5\u56de\u5f52\u5176\u672c\u8d28\uff0c\u5373\u4e3a\u4e86\u66f4\u597d\u5730\u7406\u89e3\u548c\u6bd4\u8f83\u4e0d\u540c\u6a21\u578b\u7684\u6027\u80fd\uff0c\u4e3a\u7814\u53d1\u8005\u548c\u7ec8\u7aef\u7528\u6237\u63d0\u4f9b\u53cd\u9988\uff0c\u800c\u4e0d\u662f\u4e3a\u4e86\u4ea7\u54c1\u5382\u5546\u7684\u77ed\u671f\u5229\u76ca\u3002<\/p>\n

\u65e2\u7136\u8981\u5f53\u88c1\u5224\u5458\uff0c\u8fd8\u662f\u8981\u5c3d\u91cf\u505a\u5230\u72ec\u7acb\u3001\u5ba2\u89c2\u3001\u7b2c\u4e09\u65b9\u3002<\/p>\n

\u53c2\u8003\u8d44\u6599\uff1a<\/p>\n

\u5fae\u8f6f\u7814\u7a76\u9662\u300aA Survey on Evaluation of Large Language Models\u300b<\/p>\n

\u673a\u5668\u4e4b\u5fc3\u300a13948\u9053\u9898\u76ee\uff0c\u6db5\u76d6\u5fae\u79ef\u5206\u3001\u7ebf\u4ee3\u7b4952\u4e2a\u5b66\u79d1\uff0c\u4e0a\u4ea4\u6e05\u534e\u7ed9\u4e2d\u6587\u5927\u6a21\u578b\u505a\u4e86\u4e2a\u6d4b\u8bd5\u96c6\u300b<\/p>\n

<\/p>\n

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

\u5f53\u65b0\u8d5b\u9053\u6324\u6ee1\u4e86\u6469\u62f3\u64e6\u638c\u7684\u9009\u624b\uff0c\u573a\u8fb9\u7684\u88c1\u5224\u5458\u4e5f\u5e94\u8fd0\u800c\u751f\u3002 5\u6708\uff0c\u56fd\u5bb6\u79d1...<\/p>\n","protected":false},"author":1,"featured_media":831,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[19],"tags":[],"class_list":["post-2433","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-dongtai"],"_links":{"self":[{"href":"http:\/\/www.idea2003.com\/wp-json\/wp\/v2\/posts\/2433","targetHints":{"allow":["GET"]}}],"collection":[{"href":"http:\/\/www.idea2003.com\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/www.idea2003.com\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"http:\/\/www.idea2003.com\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"http:\/\/www.idea2003.com\/wp-json\/wp\/v2\/comments?post=2433"}],"version-history":[{"count":0,"href":"http:\/\/www.idea2003.com\/wp-json\/wp\/v2\/posts\/2433\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"http:\/\/www.idea2003.com\/wp-json\/wp\/v2\/media\/831"}],"wp:attachment":[{"href":"http:\/\/www.idea2003.com\/wp-json\/wp\/v2\/media?parent=2433"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/www.idea2003.com\/wp-json\/wp\/v2\/categories?post=2433"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/www.idea2003.com\/wp-json\/wp\/v2\/tags?post=2433"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}