{"id":18245,"date":"2024-02-04T18:02:20","date_gmt":"2024-02-04T10:02:20","guid":{"rendered":"https:\/\/aif.amtbbs.org\/?p=18245"},"modified":"2024-02-04T18:02:20","modified_gmt":"2024-02-04T10:02:20","slug":"%e4%b8%80%e6%96%87%e6%90%9e%e6%87%82%e4%bd%bf%e7%94%a8-arthur-bench-%e8%bf%9b%e8%a1%8c-llm-%e8%af%84%e4%bc%b0","status":"publish","type":"post","link":"https:\/\/aif.amtbbs.org\/index.php\/2024\/02\/04\/18245\/","title":{"rendered":"\u4e00\u6587\u641e\u61c2\u4f7f\u7528 Arthur Bench \u8fdb\u884c LLM \u8bc4\u4f30"},"content":{"rendered":"<div class=\"article-desc\">\u968f\u7740 LLM \u7684\u5feb\u901f\u53d1\u5c55\u548c\u6539\u8fdb\uff0c\u6211\u4eec\u6b63\u5728\u9762\u5bf9\u65b0\u7684\u6311\u6218\u548c\u673a\u9047\u3002LLM \u7684\u80fd\u529b\u548c\u8868\u73b0\u6c34\u5e73\u4e0d\u65ad\u63d0\u9ad8\uff0c\u8fd9\u4f7f\u5f97\u57fa\u4e8e\u5355\u8bcd\u51fa\u73b0\u7684\u8bc4\u4f30\u65b9\u6cd5\u53ef\u80fd\u65e0\u6cd5\u5b8c\u5168\u6355\u6349\u5230 LLM \u751f\u6210\u6587\u672c\u7684\u8d28\u91cf\u548c\u8bed\u4e49\u51c6\u786e\u6027\u3002<\/div>\n<div id=\"postspictures\" class=\"article-content\">\n<div id=\"container\" class=\"container am-engine\" data-v-1d7a5742=\"\" data-element=\"root\">\n<p>Hello folks\uff0c\u6211\u662f Luga\uff0c\u4eca\u5929\u6211\u4eec\u6765\u804a\u4e00\u4e0b\u4eba\u5de5\u667a\u80fd(AI)\u751f\u6001\u9886\u57df\u76f8\u5173\u7684\u6280\u672f &#8211; LLM \u8bc4\u4f30 \u3002<\/p>\n<p><img data-dominant-color=\"635b7b\" data-has-transparency=\"false\" style=\"--dominant-color: #635b7b;\" loading=\"lazy\" decoding=\"async\" class=\"not-transparent alignnone size-full wp-image-18246\" src=\"https:\/\/aiforumimage.oss-cn-shanghai.aliyuncs.com\/wp-content\/uploads\/2024\/02\/6749315718cddf26b081003ffaf2bf563b9e65.jpg\" width=\"600\" height=\"399\" srcset=\"https:\/\/aiforumimage.oss-cn-shanghai.aliyuncs.com\/wp-content\/uploads\/2024\/02\/6749315718cddf26b081003ffaf2bf563b9e65.jpg 600w, https:\/\/aiforumimage.oss-cn-shanghai.aliyuncs.com\/wp-content\/uploads\/2024\/02\/6749315718cddf26b081003ffaf2bf563b9e65-300x200.jpg 300w, https:\/\/aiforumimage.oss-cn-shanghai.aliyuncs.com\/wp-content\/uploads\/2024\/02\/6749315718cddf26b081003ffaf2bf563b9e65-150x100.jpg 150w\" sizes=\"auto, (max-width: 600px) 100vw, 600px\" \/><\/p>\n<h3>\u4e00\u3001\u4f20\u7edf\u6587\u672c\u8bc4\u4f30\u9762\u4e34\u7684\u6311\u6218<\/h3>\n<p>\u8fd1\u5e74\u6765\uff0c\u968f\u7740\u5927\u578b\u8bed\u8a00\u6a21\u578b(LLM)\u7684\u5feb\u901f\u53d1\u5c55\u548c\u6539\u8fdb\uff0c\u4f20\u7edf\u7684\u6587\u672c\u8bc4\u4f30\u65b9\u6cd5\u5728\u67d0\u4e9b\u65b9\u9762\u53ef\u80fd\u5df2\u7ecf\u4e0d\u518d\u9002\u7528\u3002\u5728\u6587\u672c\u8bc4\u4f30\u9886\u57df\uff0c\u6211\u4eec\u53ef\u80fd\u5df2\u7ecf\u542c\u8bf4\u8fc7\u4e00\u4e9b\u65b9\u6cd5\uff0c\u4f8b\u5982\u57fa\u4e8e\u201c\u5355\u8bcd\u51fa\u73b0\u201d\u7684\u8bc4\u4f30\u65b9\u6cd5\uff0c\u6bd4\u5982 BLEU\uff0c\u4ee5\u53ca\u57fa\u4e8e\u201c\u9884\u8bad\u7ec3\u7684\u81ea\u7136\u8bed\u8a00\u5904\u7406\u6a21\u578b\u201d\u7684\u8bc4\u4f30\u65b9\u6cd5\uff0c\u6bd4\u5982 BERTScore\u3002<\/p>\n<p>\u5c3d\u7ba1\u8fd9\u4e9b\u65b9\u6cd5\u5728\u8fc7\u53bb\u4e00\u76f4\u975e\u5e38\u51fa\u8272\uff0c\u4f46\u968f\u7740 LLM \u7684\u751f\u6001\u6280\u672f\u7684\u4e0d\u65ad\u53d1\u5c55\uff0c\u5b83\u4eec\u663e\u5f97\u6709\u70b9\u529b\u4e0d\u4ece\u5fc3\uff0c\u65e0\u6cd5\u5b8c\u5168\u6ee1\u8db3\u5f53\u524d\u7684\u9700\u6c42\u3002<\/p>\n<p>\u968f\u7740 LLM \u7684\u5feb\u901f\u53d1\u5c55\u548c\u6539\u8fdb\uff0c\u6211\u4eec\u6b63\u5728\u9762\u5bf9\u65b0\u7684\u6311\u6218\u548c\u673a\u9047\u3002LLM \u7684\u80fd\u529b\u548c\u8868\u73b0\u6c34\u5e73\u4e0d\u65ad\u63d0\u9ad8\uff0c\u8fd9\u4f7f\u5f97\u57fa\u4e8e\u5355\u8bcd\u51fa\u73b0\u7684\u8bc4\u4f30\u65b9\u6cd5(\u5982 BLEU)\u53ef\u80fd\u65e0\u6cd5\u5b8c\u5168\u6355\u6349\u5230 LLM \u751f\u6210\u6587\u672c\u7684\u8d28\u91cf\u548c\u8bed\u4e49\u51c6\u786e\u6027\u3002LLM \u80fd\u591f\u751f\u6210\u66f4\u52a0\u6d41\u7545\u3001\u8fde\u8d2f\u4e14\u8bed\u4e49\u4e30\u5bcc\u7684\u6587\u672c\uff0c\u800c\u4f20\u7edf\u7684\u57fa\u4e8e\u5355\u8bcd\u51fa\u73b0\u7684\u8bc4\u4f30\u65b9\u6cd5\u5219\u65e0\u6cd5\u51c6\u786e\u8861\u91cf\u8fd9\u4e9b\u65b9\u9762\u7684\u4f18\u52bf\u3002<\/p>\n<p>\u6b64\u5916\uff0c\u57fa\u4e8e\u9884\u8bad\u7ec3\u6a21\u578b\u7684\u8bc4\u4f30\u65b9\u6cd5(\u5982 BERTScore)\u4e5f\u9762\u4e34\u4e00\u4e9b\u6311\u6218\u3002\u5c3d\u7ba1\u9884\u8bad\u7ec3\u6a21\u578b\u5728\u8bb8\u591a\u4efb\u52a1\u4e0a\u8868\u73b0\u51fa\u8272\uff0c\u4f46\u5b83\u4eec\u53ef\u80fd\u65e0\u6cd5\u5145\u5206\u8003\u8651\u5230 LLM \u7684\u72ec\u7279\u7279\u5f81\u4ee5\u53ca\u5176\u5728\u7279\u5b9a\u4efb\u52a1\u4e0a\u7684\u8868\u73b0\u3002LLM \u5728\u5904\u7406\u7279\u5b9a\u4efb\u52a1\u65f6\u53ef\u80fd\u4f1a\u5c55\u73b0\u51fa\u4e0e\u9884\u8bad\u7ec3\u6a21\u578b\u4e0d\u540c\u7684\u884c\u4e3a\u548c\u6027\u80fd\uff0c\u56e0\u6b64\u4ec5\u4ec5\u4f9d\u8d56\u57fa\u4e8e\u9884\u8bad\u7ec3\u6a21\u578b\u7684\u8bc4\u4f30\u65b9\u6cd5\u53ef\u80fd\u65e0\u6cd5\u5168\u9762\u8bc4\u4f30 LLM \u7684\u80fd\u529b\u3002<\/p>\n<h3>\u4e8c\u3001\u4e3a\u4ec0\u4e48\u9700\u8981 LLM \u6307\u5bfc\u8bc4\u4f30?\u4ee5\u53ca\u5e26\u6765\u7684\u6311\u6218 ?<\/h3>\n<p>\u901a\u5e38\u6765\u8bb2\uff0c\u5728\u5b9e\u9645\u7684\u4e1a\u52a1\u573a\u666f\u4e2d\uff0c\u91c7\u7528 LLM \u6307\u5bfc\u8bc4\u4f30\u8fd9\u79cd\u65b9\u6cd5\u6700\u4e3a\u4ef7\u503c\u7684\u5730\u65b9\u4e3b\u8981\u5728\u4e8e\u201c\u901f\u5ea6\u201d\u548c\u201c\u7075\u654f\u5ea6\u201d\u3002<\/p>\n<h4>1\u3001\u9ad8\u6548<\/h4>\n<p>\u9996\u5148\uff0c\u901a\u5e38\u6765\u8bf4\uff0c\u5b9e\u65bd\u901f\u5ea6\u66f4\u5feb\u3002\u76f8\u6bd4\u4e8e\u4ee5\u524d\u7684\u8bc4\u4f30\u7ba1\u9053\u6240\u9700\u7684\u5de5\u4f5c\u91cf\uff0c\u521b\u5efa LLM \u6307\u5bfc\u8bc4\u4f30\u7684\u9996\u6b21\u5b9e\u65bd\u76f8\u5bf9\u8f83\u5feb\u4e14\u5bb9\u6613\u3002\u5bf9\u4e8e LLM \u6307\u5bfc\u7684\u8bc4\u4f30\uff0c\u6211\u4eec\u53ea\u9700\u8981\u51c6\u5907\u4e24\u4ef6\u4e8b\u60c5\uff1a\u7528\u6587\u5b57\u63cf\u8ff0\u8bc4\u4f30\u6807\u51c6\uff0c\u5e76\u63d0\u4f9b\u4e00\u4e9b\u5728\u63d0\u793a\u6a21\u677f\u4e2d\u4f7f\u7528\u7684\u793a\u4f8b\u3002\u76f8\u5bf9\u4e8e\u6784\u5efa\u81ea\u5df1\u7684\u9884\u8bad\u7ec3 NLP \u6a21\u578b(\u6216\u5fae\u8c03\u73b0\u6709\u7684 NLP \u6a21\u578b)\u4ee5\u7528\u4f5c\u8bc4\u4f30\u5668\u6240\u9700\u7684\u5de5\u4f5c\u91cf\u548c\u6570\u636e\u6536\u96c6\u91cf\uff0c\u4f7f\u7528 LLM \u6765\u5b8c\u6210\u8fd9\u4e9b\u4efb\u52a1\u66f4\u4e3a\u9ad8\u6548\u3002\u4f7f\u7528 LLM\uff0c\u8bc4\u4f30\u6807\u51c6\u7684\u8fed\u4ee3\u901f\u5ea6\u8981\u5feb\u5f97\u591a\u3002<\/p>\n<h4>2\u3001\u654f\u611f\u6027<\/h4>\n<p>\u5176\u6b21\uff0cLLM \u901a\u5e38\u66f4\u52a0\u654f\u611f\u3002\u8fd9\u79cd\u654f\u611f\u6027\u53ef\u80fd\u5e26\u6765\u79ef\u6781\u7684\u65b9\u9762\uff0c\u4e0e\u9884\u8bad\u7ec3\u7684 NLP \u6a21\u578b\u548c\u4e4b\u524d\u8ba8\u8bba\u7684\u8bc4\u4f30\u65b9\u6cd5\u76f8\u6bd4\uff0cLLM \u66f4\u80fd\u7075\u6d3b\u5730\u5904\u7406\u8fd9\u4e9b\u60c5\u51b5\u3002\u7136\u800c\uff0c\u8fd9\u79cd\u654f\u611f\u6027\u4e5f\u53ef\u80fd\u5bfc\u81f4 LLM \u7684\u8bc4\u4f30\u7ed3\u679c\u53d8\u5f97\u975e\u5e38\u4e0d\u53ef\u9884\u6d4b\u3002<\/p>\n<p>\u6b63\u5982\u6211\u4eec\u4e4b\u524d\u8ba8\u8bba\u7684\u90a3\u6837\uff0c\u4e0e\u5176\u4ed6\u8bc4\u4f30\u65b9\u6cd5\u76f8\u6bd4\uff0cLLM \u8bc4\u4f30\u8005\u66f4\u52a0\u654f\u611f\u3002\u5c06 LLM \u4f5c\u4e3a\u8bc4\u4f30\u5668\u6709\u8bb8\u591a\u4e0d\u540c\u7684\u914d\u7f6e\u65b9\u6cd5\uff0c\u6839\u636e\u6240\u9009\u62e9\u7684\u914d\u7f6e\uff0c\u5176\u884c\u4e3a\u53ef\u80fd\u4f1a\u6709\u5f88\u5927\u7684\u5dee\u5f02\u3002\u540c\u65f6\uff0c\u53e6\u4e00\u4e2a\u6311\u6218\u5728\u4e8e\uff0c\u5982\u679c\u8bc4\u4f30\u6d89\u53ca\u592a\u591a\u7684\u63a8\u7406\u6b65\u9aa4\u6216\u9700\u8981\u540c\u65f6\u5904\u7406\u592a\u591a\u7684\u53d8\u91cf\uff0cLLM \u8bc4\u4f30\u8005\u53ef\u80fd\u4f1a\u9677\u5165\u56f0\u5883\u3002<\/p>\n<p>\u7531\u4e8e LLM \u7684\u7279\u6027\uff0c\u5176\u8bc4\u4f30\u7ed3\u679c\u53ef\u80fd\u4f1a\u53d7\u5230\u4e0d\u540c\u914d\u7f6e\u548c\u53c2\u6570\u8bbe\u7f6e\u7684\u5f71\u54cd\u3002\u8fd9\u610f\u5473\u7740\u5bf9 LLM \u8fdb\u884c\u8bc4\u4f30\u65f6\uff0c\u9700\u8981\u4ed4\u7ec6\u9009\u62e9\u548c\u914d\u7f6e\u6a21\u578b\uff0c\u4ee5\u786e\u4fdd\u5176\u884c\u4e3a\u7b26\u5408\u9884\u671f\u3002\u4e0d\u540c\u7684\u914d\u7f6e\u53ef\u80fd\u5bfc\u81f4\u4e0d\u540c\u7684\u8f93\u51fa\u7ed3\u679c\uff0c\u56e0\u6b64\u8bc4\u4f30\u8005\u9700\u8981\u82b1\u8d39\u4e00\u5b9a\u7684\u65f6\u95f4\u548c\u7cbe\u529b\u6765\u8c03\u6574\u548c\u4f18\u5316 LLM \u7684\u8bbe\u7f6e\uff0c\u4ee5\u83b7\u5f97\u51c6\u786e\u548c\u53ef\u9760\u7684\u8bc4\u4f30\u7ed3\u679c\u3002<\/p>\n<p>\u6b64\u5916\uff0c\u5f53\u9762\u5bf9\u9700\u8981\u8fdb\u884c\u590d\u6742\u63a8\u7406\u6216\u540c\u65f6\u5904\u7406\u591a\u4e2a\u53d8\u91cf\u7684\u8bc4\u4f30\u4efb\u52a1\u65f6\uff0c\u8bc4\u4f30\u8005\u53ef\u80fd\u4f1a\u9762\u4e34\u4e00\u4e9b\u6311\u6218\u3002\u8fd9\u662f\u56e0\u4e3a LLM \u7684\u63a8\u7406\u80fd\u529b\u5728\u5904\u7406\u590d\u6742\u60c5\u5883\u65f6\u53ef\u80fd\u53d7\u9650\u3002LLM \u53ef\u80fd\u9700\u8981\u8fdb\u884c\u66f4\u591a\u7684\u52aa\u529b\u6765\u89e3\u51b3\u8fd9\u4e9b\u4efb\u52a1\uff0c\u4ee5\u786e\u4fdd\u8bc4\u4f30\u7684\u51c6\u786e\u6027\u548c\u53ef\u9760\u6027\u3002<\/p>\n<h3>\u4e09\u3001\u4ec0\u4e48\u662f Arthur Bench ?<\/h3>\n<p>Arthur Bench \u662f\u4e00\u4e2a\u5f00\u6e90\u7684\u8bc4\u4f30\u5de5\u5177\uff0c\u7528\u4e8e\u6bd4\u8f83\u751f\u6210\u6587\u672c\u6a21\u578b (LLM) \u7684\u6027\u80fd\u3002\u5b83\u53ef\u4ee5\u7528\u4e8e\u8bc4\u4f30\u4e0d\u540c LLM \u6a21\u578b\u3001\u63d0\u793a\u548c\u8d85\u53c2\u6570\uff0c\u5e76\u63d0\u4f9b\u6709\u5173 LLM \u5728\u5404\u79cd\u4efb\u52a1\u4e0a\u7684\u6027\u80fd\u7684\u8be6\u7ec6\u62a5\u544a\u3002<\/p>\n<p>Arthur Bench \u7684\u4e3b\u8981\u529f\u80fd\u5305\u62ec\uff1aArthur Bench \u7684\u4e3b\u8981\u529f\u80fd\u5305\u62ec\uff1a<\/p>\n<ul data-id=\"u738a58b-vrsueohA\">\n<li data-id=\"ld70c578-pGV7EJsc\">\u6bd4\u8f83\u4e0d\u540c LLM \u6a21\u578b\uff1aArthur Bench \u53ef\u4ee5\u7528\u4e8e\u6bd4\u8f83\u4e0d\u540c LLM \u6a21\u578b\u7684\u6027\u80fd\uff0c\u5305\u62ec\u6765\u81ea\u4e0d\u540c\u4f9b\u5e94\u5546\u7684\u6a21\u578b\u3001\u4e0d\u540c\u7248\u672c\u7684\u6a21\u578b\u4ee5\u53ca\u4f7f\u7528\u4e0d\u540c\u8bad\u7ec3\u6570\u636e\u96c6\u7684\u6a21\u578b\u3002<\/li>\n<li data-id=\"ld70c578-GDcpYbzU\">\u8bc4\u4f30\u63d0\u793a\uff1aArthur Bench \u53ef\u4ee5\u7528\u4e8e\u8bc4\u4f30\u4e0d\u540c\u63d0\u793a\u5bf9 LLM \u6027\u80fd\u7684\u5f71\u54cd\u3002\u63d0\u793a\u662f\u7528\u4e8e\u6307\u5bfc LLM \u751f\u6210\u6587\u672c\u7684\u6307\u4ee4\u3002<\/li>\n<li data-id=\"ld70c578-tatS8jem\">\u6d4b\u8bd5\u8d85\u53c2\u6570\uff1aArthur Bench \u53ef\u4ee5\u7528\u4e8e\u6d4b\u8bd5\u4e0d\u540c\u8d85\u53c2\u6570\u5bf9 LLM \u6027\u80fd\u7684\u5f71\u54cd\u3002\u8d85\u53c2\u6570\u662f\u63a7\u5236 LLM \u884c\u4e3a\u7684\u8bbe\u7f6e\u3002<\/li>\n<\/ul>\n<p>\u901a\u5e38\u800c\u8a00\uff0cArthur Bench \u5de5\u4f5c\u6d41\u7a0b\u4e3b\u8981\u6d89\u53ca\u5982\u4e0b\u9636\u6bb5\uff0c\u5177\u4f53\u8a73\u7d30\u89e3\u6790\u5982\u4e0b\u6240\u793a\uff1a<\/p>\n<p><img data-dominant-color=\"fafafb\" data-has-transparency=\"false\" style=\"--dominant-color: #fafafb;\" loading=\"lazy\" decoding=\"async\" class=\"not-transparent alignnone size-full wp-image-18247\" src=\"https:\/\/aiforumimage.oss-cn-shanghai.aliyuncs.com\/wp-content\/uploads\/2024\/02\/13eb1a7174fa7c7c26651582c01415723b5e0f.png\" width=\"1080\" height=\"413\" alt=\"\" title=\"\u56fe\u7247\" srcset=\"https:\/\/aiforumimage.oss-cn-shanghai.aliyuncs.com\/wp-content\/uploads\/2024\/02\/13eb1a7174fa7c7c26651582c01415723b5e0f.png 1080w, https:\/\/aiforumimage.oss-cn-shanghai.aliyuncs.com\/wp-content\/uploads\/2024\/02\/13eb1a7174fa7c7c26651582c01415723b5e0f-300x115.png 300w, https:\/\/aiforumimage.oss-cn-shanghai.aliyuncs.com\/wp-content\/uploads\/2024\/02\/13eb1a7174fa7c7c26651582c01415723b5e0f-1024x392.png 1024w, https:\/\/aiforumimage.oss-cn-shanghai.aliyuncs.com\/wp-content\/uploads\/2024\/02\/13eb1a7174fa7c7c26651582c01415723b5e0f-150x57.png 150w, https:\/\/aiforumimage.oss-cn-shanghai.aliyuncs.com\/wp-content\/uploads\/2024\/02\/13eb1a7174fa7c7c26651582c01415723b5e0f-768x294.png 768w\" sizes=\"auto, (max-width: 1080px) 100vw, 1080px\" \/><\/p>\n<h4>1. \u4efb\u52a1\u5b9a\u4e49<\/h4>\n<p>\u5728\u6b64\u9636\u6bb5\uff0c\u6211\u4eec\u9700\u8981\u660e\u786e\u6211\u4eec\u7684\u8bc4\u4f30\u76ee\u6807\uff0cArthur Bench \u652f\u6301\u591a\u79cd\u8bc4\u4f30\u4efb\u52a1\uff0c\u5305\u62ec\uff1a<\/p>\n<ul data-id=\"u738a58b-tQvcYnaz\">\n<li data-id=\"ld70c578-hRoWVEBg\">\u95ee\u7b54\uff1a\u6d4b\u8bd5 LLM \u5bf9\u5f00\u653e\u5f0f\u3001\u6311\u6218\u6027\u6216\u591a\u4e49\u6027\u95ee\u9898\u7684\u7406\u89e3\u548c\u56de\u7b54\u80fd\u529b\u3002<\/li>\n<li data-id=\"ld70c578-Suko53Ks\">\u6458\u8981\uff1a\u8bc4\u4f30 LLM \u63d0\u53d6\u6587\u672c\u5173\u952e\u4fe1\u606f\u5e76\u751f\u6210\u7b80\u6d01\u6458\u8981\u7684\u80fd\u529b\u3002<\/li>\n<li data-id=\"ld70c578-en3ly2XP\">\u7ffb\u8bd1\uff1a\u8003\u5bdf LLM \u5728\u4e0d\u540c\u8bed\u8a00\u4e4b\u95f4\u8fdb\u884c\u51c6\u786e\u3001\u6d41\u7545\u7ffb\u8bd1\u7684\u80fd\u529b\u3002<\/li>\n<li data-id=\"ld70c578-ULNnRbAM\">\u4ee3\u7801\u751f\u6210\uff1a\u6d4b\u8bd5 LLM \u6839\u636e\u81ea\u7136\u8bed\u8a00\u63cf\u8ff0\u751f\u6210\u4ee3\u7801\u7684\u80fd\u529b\u3002<\/li>\n<\/ul>\n<h4>2. \u6a21\u578b\u9009\u62e9<\/h4>\n<p>\u5728\u6b64\u9636\u6bb5\uff0c\u4e3b\u8981\u5de5\u4f5c\u4e3a\u7b5b\u9009\u8bc4\u4f30\u5bf9\u8c61\u3002Arthur Bench \u652f\u6301\u591a\u79cd LLM \u6a21\u578b\uff0c\u6db5\u76d6\u6765\u81ea OpenAI\u3001Google AI\u3001Microsoft \u7b49\u77e5\u540d\u673a\u6784\u7684\u9886\u5148\u6280\u672f\uff0c\u5982 GPT-3\u3001LaMDA\u3001Megatron-Turing NLG \u7b49\u3002\u6211\u4eec\u53ef\u4ee5\u6839\u636e\u7814\u7a76\u9700\u6c42\u9009\u62e9\u7279\u5b9a\u6a21\u578b\u8fdb\u884c\u8bc4\u4f30\u3002<\/p>\n<h4>3. \u53c2\u6570\u914d\u7f6e<\/h4>\n<p>\u5b8c\u6210\u6a21\u578b\u9009\u62e9\u540e\uff0c\u63a5\u4e0b\u6765\u8fdb\u884c\u7cbe\u7ec6\u5316\u8c03\u63a7\u5de5\u4f5c\u3002\u4e3a\u4e86\u66f4\u7cbe\u51c6\u5730\u8bc4\u4f30 LLM \u6027\u80fd\uff0cArthur Bench \u5141\u8bb8\u7528\u6237\u914d\u7f6e\u63d0\u793a\u548c\u8d85\u53c2\u6570\u3002<\/p>\n<ul data-id=\"u738a58b-cAnfy6cZ\">\n<li data-id=\"ld70c578-sOJDtAXf\">\u63d0\u793a\uff1a\u6307\u5f15 LLM \u751f\u6210\u6587\u672c\u7684\u65b9\u5411\u548c\u5185\u5bb9\uff0c\u4f8b\u5982\u95ee\u9898\u3001\u63cf\u8ff0\u6216\u6307\u4ee4\u3002<\/li>\n<li data-id=\"ld70c578-ECFqf8yI\">\u8d85\u53c2\u6570\uff1a\u63a7\u5236 LLM \u884c\u4e3a\u7684\u5173\u952e\u8bbe\u7f6e\uff0c\u4f8b\u5982\u5b66\u4e60\u7387\u3001\u8bad\u7ec3\u6b65\u6570\u3001\u6a21\u578b\u67b6\u6784\u7b49\u3002<\/li>\n<\/ul>\n<p>\u901a\u8fc7\u7cbe\u7ec6\u5316\u914d\u7f6e\uff0c\u6211\u4eec\u53ef\u4ee5\u6df1\u5165\u63a2\u7d22 LLM \u5728\u4e0d\u540c\u53c2\u6570\u8bbe\u7f6e\u4e0b\u7684\u8868\u73b0\u5dee\u5f02\uff0c\u83b7\u5f97\u66f4\u5177\u53c2\u8003\u4ef7\u503c\u7684\u8bc4\u4f30\u7ed3\u679c\u3002<\/p>\n<h4>4. \u8bc4\u4f30\u8fd0\u884c\uff1a\u81ea\u52a8\u5316\u6d41\u7a0b<\/h4>\n<p>\u6700\u540e\u4e00\u6b65\uff0c\u5373\u501f\u52a9\u81ea\u52a8\u5316\u6d41\u7a0b\u8fdb\u884c\u4efb\u52a1\u8bc4\u4f30\u3002\u901a\u5e38\u60c5\u51b5\u4e0b\uff0cArthur Bench \u63d0\u4f9b\u81ea\u52a8\u5316\u8bc4\u4f30\u6d41\u7a0b\uff0c\u53ea\u9700\u7b80\u5355\u914d\u7f6e\u5373\u53ef\u8fd0\u884c\u8bc4\u4f30\u4efb\u52a1\u3002\u5b83\u5c06\u81ea\u52a8\u6267\u884c\u4ee5\u4e0b\u6b65\u9aa4\uff1a<\/p>\n<ul data-id=\"u738a58b-X1lHmaRV\">\n<li data-id=\"ld70c578-AQvyIyUc\">\u8c03\u7528 LLM \u6a21\u578b\u5e76\u751f\u6210\u6587\u672c\u8f93\u51fa\u3002<\/li>\n<li data-id=\"ld70c578-QBnyjtJi\">\u9488\u5bf9\u7279\u5b9a\u4efb\u52a1\uff0c\u5e94\u7528\u76f8\u5e94\u7684\u8bc4\u4f30\u6307\u6807\u8fdb\u884c\u5206\u6790\u3002<\/li>\n<li data-id=\"ld70c578-RJSEepzc\">\u751f\u6210\u8be6\u7ec6\u62a5\u544a\uff0c\u5448\u73b0\u8bc4\u4f30\u7ed3\u679c\u3002<\/li>\n<\/ul>\n<h3>\u56db\u3001Arthur Bench \u4f7f\u7528\u573a\u666f\u5206\u6790<\/h3>\n<p>\u4f5c\u4e3a\u4e00\u79cd\u5feb\u901f\u3001\u6570\u636e\u9a71\u52a8\u7684 LLM \u8bc4\u4f30\u7684\u5173\u952e\uff0cArthur Bench \u4e3b\u8981\u63d0\u4f9b\u5982\u4e0b\u89e3\u51b3\u65b9\u6848\uff0c\u5177\u4f53\u6d89\u53ca\uff1a<\/p>\n<h4>1\u3001\u6a21\u578b\u9009\u62e9\u548c\u9a8c\u8bc1<\/h4>\n<p>\u6a21\u578b\u9009\u62e9\u548c\u9a8c\u8bc1\u662f\u4eba\u5de5\u667a\u80fd\u9886\u57df\u4e2d\u81f3\u5173\u91cd\u8981\u7684\u5173\u952e\u6b65\u9aa4\uff0c\u5bf9\u4e8e\u786e\u4fdd\u6a21\u578b\u7684\u6709\u6548\u6027\u548c\u53ef\u9760\u6027\u5177\u6709\u91cd\u8981\u610f\u4e49\u3002\u5728\u8fd9\u4e2a\u8fc7\u7a0b\u4e2d\uff0cArthur Bench \u7684\u89d2\u8272\u975e\u5e38\u5173\u952e\u3002\u4ed6\u7684\u76ee\u6807\u662f\u4e3a\u516c\u53f8\u63d0\u4f9b\u4e00\u4e2a\u53ef\u9760\u7684\u6bd4\u8f83\u6846\u67b6\uff0c\u901a\u8fc7\u4f7f\u7528\u4e00\u81f4\u7684\u6307\u6807\u548c\u8bc4\u4f30\u65b9\u6cd5\uff0c\u5e2e\u52a9\u4ed6\u4eec\u5728\u4f17\u591a\u5927\u578b\u8bed\u8a00\u6a21\u578b(LLM)\u9009\u9879\u4e2d\u505a\u51fa\u660e\u667a\u7684\u51b3\u7b56\u3002<\/p>\n<p><img data-dominant-color=\"7a6777\" data-has-transparency=\"false\" style=\"--dominant-color: #7a6777;\" loading=\"lazy\" decoding=\"async\" class=\"not-transparent alignnone size-full wp-image-18248\" src=\"https:\/\/aiforumimage.oss-cn-shanghai.aliyuncs.com\/wp-content\/uploads\/2024\/02\/b28d64a61ac007ec5d19224ff03a36522c6eb9.png\" width=\"1080\" height=\"742\" alt=\"\" title=\"\u56fe\u7247\" srcset=\"https:\/\/aiforumimage.oss-cn-shanghai.aliyuncs.com\/wp-content\/uploads\/2024\/02\/b28d64a61ac007ec5d19224ff03a36522c6eb9.png 1080w, https:\/\/aiforumimage.oss-cn-shanghai.aliyuncs.com\/wp-content\/uploads\/2024\/02\/b28d64a61ac007ec5d19224ff03a36522c6eb9-300x206.png 300w, https:\/\/aiforumimage.oss-cn-shanghai.aliyuncs.com\/wp-content\/uploads\/2024\/02\/b28d64a61ac007ec5d19224ff03a36522c6eb9-1024x704.png 1024w, https:\/\/aiforumimage.oss-cn-shanghai.aliyuncs.com\/wp-content\/uploads\/2024\/02\/b28d64a61ac007ec5d19224ff03a36522c6eb9-150x103.png 150w, https:\/\/aiforumimage.oss-cn-shanghai.aliyuncs.com\/wp-content\/uploads\/2024\/02\/b28d64a61ac007ec5d19224ff03a36522c6eb9-768x528.png 768w\" sizes=\"auto, (max-width: 1080px) 100vw, 1080px\" \/><\/p>\n<p>Arthur Bench \u5c06\u8fd0\u7528\u4ed6\u7684\u4e13\u4e1a\u77e5\u8bc6\u548c\u7ecf\u9a8c\u6765\u8bc4\u4f30\u6bcf\u4e2a LLM \u9009\u9879\uff0c\u5e76\u786e\u4fdd\u4f7f\u7528\u4e00\u81f4\u7684\u6307\u6807\u6765\u6bd4\u8f83\u5b83\u4eec\u7684\u4f18\u52bf\u548c\u52a3\u52bf\u3002\u4ed6\u5c06\u7efc\u5408\u8003\u8651\u8bf8\u5982\u6a21\u578b\u6027\u80fd\u3001\u51c6\u786e\u6027\u3001\u901f\u5ea6\u3001\u8d44\u6e90\u9700\u6c42\u7b49\u56e0\u7d20\uff0c\u4ee5\u786e\u4fdd\u516c\u53f8\u80fd\u591f\u505a\u51fa\u660e\u667a\u800c\u660e\u786e\u7684\u9009\u62e9\u3002<\/p>\n<p>\u901a\u8fc7\u4f7f\u7528\u4e00\u81f4\u7684\u6307\u6807\u548c\u8bc4\u4f30\u65b9\u6cd5\uff0cArthur Bench \u5c06\u4e3a\u516c\u53f8\u63d0\u4f9b\u4e00\u4e2a\u53ef\u9760\u7684\u6bd4\u8f83\u6846\u67b6\uff0c\u4f7f\u4ed6\u4eec\u80fd\u591f\u5168\u9762\u8bc4\u4f30\u6bcf\u4e2a LLM \u9009\u9879\u7684\u4f18\u70b9\u548c\u5c40\u9650\u6027\u3002\u8fd9\u5c06\u4f7f\u516c\u53f8\u80fd\u591f\u505a\u51fa\u660e\u667a\u7684\u51b3\u7b56\uff0c\u4ee5\u6700\u5927\u7a0b\u5ea6\u5730\u5229\u7528\u4eba\u5de5\u667a\u80fd\u9886\u57df\u7684\u5feb\u901f\u53d1\u5c55\uff0c\u5e76\u786e\u4fdd\u4ed6\u4eec\u7684\u5e94\u7528\u7a0b\u5e8f\u80fd\u591f\u83b7\u5f97\u6700\u4f73\u7684\u4f53\u9a8c\u6548\u679c\u3002<\/p>\n<h4>2\u3001\u9884\u7b97\u548c\u9690\u79c1\u4f18\u5316<\/h4>\n<p>\u5728\u9009\u62e9\u4eba\u5de5\u667a\u80fd\u6a21\u578b\u65f6\uff0c\u5e76\u975e\u6240\u6709\u5e94\u7528\u7a0b\u5e8f\u90fd\u9700\u8981\u6700\u5148\u8fdb\u6216\u6700\u6602\u8d35\u7684\u5927\u578b\u8bed\u8a00\u6a21\u578b(LLM)\u3002\u5728\u67d0\u4e9b\u60c5\u51b5\u4e0b\uff0c\u4f7f\u7528\u6210\u672c\u66f4\u4f4e\u7684\u4eba\u5de5\u667a\u80fd\u6a21\u578b\u4e5f\u53ef\u4ee5\u6ee1\u8db3\u4efb\u52a1\u9700\u6c42\u3002<\/p>\n<p>\u8fd9\u79cd\u9884\u7b97\u4f18\u5316\u7684\u65b9\u6cd5\u53ef\u4ee5\u5e2e\u52a9\u516c\u53f8\u5728\u6709\u9650\u7684\u8d44\u6e90\u4e0b\u505a\u51fa\u660e\u667a\u7684\u9009\u62e9\u3002\u800c\u4e0d\u5fc5\u8ffd\u6c42\u6700\u6602\u8d35\u6216\u6700\u5148\u8fdb\u7684\u6a21\u578b\uff0c\u800c\u662f\u6839\u636e\u5177\u4f53\u9700\u6c42\u9009\u62e9\u5408\u9002\u7684\u6a21\u578b\u3002\u8f83\u4e3a\u7ecf\u6d4e\u5b9e\u60e0\u7684\u6a21\u578b\u53ef\u80fd\u5728\u67d0\u4e9b\u65b9\u9762\u7684\u6027\u80fd\u7565\u4f4e\u4e8e\u6700\u5148\u8fdb\u7684 LLM\uff0c\u4f46\u5bf9\u4e8e\u4e00\u4e9b\u7b80\u5355\u6216\u6807\u51c6\u7684\u4efb\u52a1\u6765\u8bf4\uff0cArthur Bench \u4ecd\u7136\u80fd\u591f\u63d0\u4f9b\u6ee1\u8db3\u9700\u6c42\u7684\u89e3\u51b3\u65b9\u6848\u3002<\/p>\n<p>\u6b64\u5916\uff0cArthur Bench \u5f3a\u8c03\u5c06\u6a21\u578b\u5f15\u5165\u5185\u90e8\u53ef\u4ee5\u66f4\u597d\u5730\u63a7\u5236\u6570\u636e\u9690\u79c1\u3002\u5bf9\u4e8e\u6d89\u53ca\u654f\u611f\u6570\u636e\u6216\u9690\u79c1\u95ee\u9898\u7684\u5e94\u7528\u7a0b\u5e8f\uff0c\u516c\u53f8\u53ef\u80fd\u66f4\u503e\u5411\u4e8e\u4f7f\u7528\u81ea\u5df1\u5185\u90e8\u8bad\u7ec3\u7684\u6a21\u578b\uff0c\u800c\u4e0d\u662f\u4f9d\u8d56\u5916\u90e8\u7684\u7b2c\u4e09\u65b9LLM\u3002\u901a\u8fc7\u4f7f\u7528\u5185\u90e8\u6a21\u578b\uff0c\u516c\u53f8\u53ef\u4ee5\u66f4\u597d\u5730\u638c\u63e1\u6570\u636e\u7684\u5904\u7406\u548c\u5b58\u50a8\uff0c\u66f4\u597d\u5730\u4fdd\u62a4\u6570\u636e\u9690\u79c1\u3002<\/p>\n<h4>3\u3001\u5c06\u5b66\u672f\u57fa\u51c6\u8f6c\u5316\u4e3a\u73b0\u5b9e\u4e16\u754c\u7684\u8868\u73b0<\/h4>\n<p>\u5b66\u672f\u57fa\u51c6\u662f\u6307\u5728\u5b66\u672f\u7814\u7a76\u4e2d\u5efa\u7acb\u7684\u6a21\u578b\u8bc4\u4f30\u6307\u6807\u548c\u65b9\u6cd5\u3002\u8fd9\u4e9b\u6307\u6807\u548c\u65b9\u6cd5\u901a\u5e38\u662f\u9488\u5bf9\u7279\u5b9a\u4efb\u52a1\u6216\u9886\u57df\u7684\uff0c\u80fd\u591f\u6709\u6548\u8bc4\u4f30\u6a21\u578b\u5728\u8be5\u4efb\u52a1\u6216\u9886\u57df\u7684\u6027\u80fd\u3002<\/p>\n<p>\u7136\u800c\uff0c\u5b66\u672f\u57fa\u51c6\u5e76\u4e0d\u603b\u662f\u80fd\u591f\u76f4\u63a5\u53cd\u6620\u6a21\u578b\u5728\u73b0\u5b9e\u4e16\u754c\u4e2d\u7684\u8868\u73b0\u3002\u8fd9\u662f\u56e0\u4e3a\u73b0\u5b9e\u4e16\u754c\u4e2d\u7684\u5e94\u7528\u573a\u666f\u5f80\u5f80\u66f4\u52a0\u590d\u6742\uff0c\u9700\u8981\u8003\u8651\u66f4\u591a\u56e0\u7d20\uff0c\u4f8b\u5982\u6570\u636e\u5206\u5e03\u3001\u6a21\u578b\u90e8\u7f72\u73af\u5883\u7b49\u3002<\/p>\n<p>Arthur Bench \u53ef\u4ee5\u5e2e\u52a9\u5c06\u5b66\u672f\u57fa\u51c6\u8f6c\u5316\u4e3a\u73b0\u5b9e\u4e16\u754c\u7684\u8868\u73b0\u3002 \u5b83\u901a\u8fc7\u4ee5\u4e0b\u65b9\u5f0f\u5b9e\u73b0\u8fd9\u4e00\u76ee\u6807\uff1a<\/p>\n<ul data-id=\"u738a58b-C2dDXAJq\">\n<li data-id=\"ld70c578-o79eeeOr\">\u63d0\u4f9b\u4e00\u7ec4\u5168\u9762\u7684\u8bc4\u4f30\u6307\u6807\uff0c \u6db5\u76d6\u6a21\u578b\u7684\u51c6\u786e\u6027\u3001\u6548\u7387\u3001\u9c81\u68d2\u6027\u7b49\u591a\u4e2a\u65b9\u9762\u3002\u8fd9\u4e9b\u6307\u6807\u4e0d\u4ec5\u80fd\u591f\u53cd\u6620\u6a21\u578b\u5728\u5b66\u672f\u57fa\u51c6\u4e0b\u7684\u8868\u73b0\uff0c\u4e5f\u80fd\u53cd\u6620\u6a21\u578b\u5728\u73b0\u5b9e\u4e16\u754c\u4e2d\u7684\u6f5c\u5728\u8868\u73b0\u3002<\/li>\n<li data-id=\"ld70c578-LEvxpaqK\">\u652f\u6301\u591a\u79cd\u6a21\u578b\u7c7b\u578b\uff0c \u80fd\u591f\u5bf9\u4e0d\u540c\u7c7b\u578b\u7684\u6a21\u578b\u8fdb\u884c\u6bd4\u8f83\u3002\u8fd9\u4f7f\u5f97\u4f01\u4e1a\u80fd\u591f\u9009\u62e9\u6700\u9002\u5408\u5176\u5e94\u7528\u573a\u666f\u7684\u6a21\u578b\u3002<\/li>\n<li data-id=\"ld70c578-5ALGXJrf\">\u63d0\u4f9b\u53ef\u89c6\u5316\u5206\u6790\u5de5\u5177\uff0c \u5e2e\u52a9\u4f01\u4e1a\u76f4\u89c2\u5730\u4e86\u89e3\u4e0d\u540c\u6a21\u578b\u7684\u8868\u73b0\u5dee\u5f02\u3002\u8fd9\u4f7f\u5f97\u4f01\u4e1a\u80fd\u591f\u66f4\u5bb9\u6613\u5730\u505a\u51fa\u51b3\u7b56\u3002<\/li>\n<\/ul>\n<h3>\u4e94\u3001Arthur Bench \u7279\u6027\u5206\u6790<\/h3>\n<p>\u4f5c\u4e3a\u4e00\u79cd\u5feb\u901f\u3001\u6570\u636e\u9a71\u52a8\u7684 LLM \u8bc4\u4f30\u7684\u5173\u952e\uff0cArthur Bench \u5177\u6709\u5982\u4e0b\u7279\u6027\uff1a<\/p>\n<h4>1\u3001\u5168\u5957\u8bc4\u5206\u6307\u6807<\/h4>\n<p>Arthur Bench \u62e5\u6709\u4e00\u6574\u5957\u8bc4\u5206\u6307\u6807\uff0c\u6db5\u76d6\u4e86\u4ece\u603b\u7ed3\u8d28\u91cf\u5230\u7528\u6237\u4f53\u9a8c\u7684\u5404\u4e2a\u65b9\u9762\u3002\u4ed6\u53ef\u4ee5\u968f\u65f6\u5229\u7528\u8fd9\u4e9b\u8bc4\u5206\u6307\u6807\u6765\u5bf9\u4e0d\u540c\u7684\u6a21\u578b\u8fdb\u884c\u8bc4\u4f30\u548c\u6bd4\u8f83\u3002\u8fd9\u4e9b\u8bc4\u5206\u6307\u6807\u7684\u7efc\u5408\u8fd0\u7528\u53ef\u4ee5\u5e2e\u52a9\u4ed6\u5168\u9762\u4e86\u89e3\u6bcf\u4e2a\u6a21\u578b\u7684\u4f18\u52bf\u548c\u52a3\u52bf\u3002<\/p>\n<p>\u8fd9\u4e9b\u8bc4\u5206\u6307\u6807\u7684\u8303\u56f4\u975e\u5e38\u5e7f\u6cdb\uff0c\u5305\u62ec\u4f46\u4e0d\u9650\u4e8e\u603b\u7ed3\u8d28\u91cf\u3001\u51c6\u786e\u6027\u3001\u6d41\u7545\u6027\u3001\u8bed\u6cd5\u6b63\u786e\u6027\u3001\u4e0a\u4e0b\u6587\u7406\u89e3\u80fd\u529b\u3001\u903b\u8f91\u8fde\u8d2f\u6027\u7b49\u3002Arthur Bench \u5c06\u6839\u636e\u8fd9\u4e9b\u6307\u6807\u5bf9\u6bcf\u4e2a\u6a21\u578b\u8fdb\u884c\u8bc4\u4f30\uff0c\u5e76\u5c06\u7ed3\u679c\u6574\u5408\u4e3a\u4e00\u4e2a\u7efc\u5408\u8bc4\u5206\uff0c\u4ee5\u8f85\u52a9\u516c\u53f8\u505a\u51fa\u660e\u667a\u7684\u51b3\u7b56\u3002<\/p>\n<p>\u6b64\u5916\uff0c\u5982\u679c\u516c\u53f8\u6709\u7279\u5b9a\u7684\u9700\u6c42\u6216\u5173\u6ce8\u70b9\uff0cArthur Bench \u8fd8\u53ef\u4ee5\u6839\u636e\u516c\u53f8\u7684\u8981\u6c42\u521b\u5efa\u548c\u6dfb\u52a0\u81ea\u5b9a\u4e49\u7684\u8bc4\u5206\u6307\u6807\u3002\u8fd9\u6837\u4ee5\u4fbf\u80fd\u591f\u66f4\u597d\u5730\u6ee1\u8db3\u516c\u53f8\u7684\u5177\u4f53\u9700\u6c42\uff0c\u5e76\u786e\u4fdd\u8bc4\u4f30\u8fc7\u7a0b\u4e0e\u516c\u53f8\u7684\u76ee\u6807\u548c\u6807\u51c6\u76f8\u7b26\u5408\u3002<\/p>\n<p><img data-dominant-color=\"edeaec\" data-has-transparency=\"false\" style=\"--dominant-color: #edeaec;\" loading=\"lazy\" decoding=\"async\" class=\"not-transparent alignnone size-full wp-image-18249\" src=\"https:\/\/aiforumimage.oss-cn-shanghai.aliyuncs.com\/wp-content\/uploads\/2024\/02\/d5c0a9622cb6499cd30471beece1babc095d92.png\" width=\"1024\" height=\"594\" alt=\"\" title=\"\u56fe\u7247\" srcset=\"https:\/\/aiforumimage.oss-cn-shanghai.aliyuncs.com\/wp-content\/uploads\/2024\/02\/d5c0a9622cb6499cd30471beece1babc095d92.png 1024w, https:\/\/aiforumimage.oss-cn-shanghai.aliyuncs.com\/wp-content\/uploads\/2024\/02\/d5c0a9622cb6499cd30471beece1babc095d92-300x174.png 300w, https:\/\/aiforumimage.oss-cn-shanghai.aliyuncs.com\/wp-content\/uploads\/2024\/02\/d5c0a9622cb6499cd30471beece1babc095d92-150x87.png 150w, https:\/\/aiforumimage.oss-cn-shanghai.aliyuncs.com\/wp-content\/uploads\/2024\/02\/d5c0a9622cb6499cd30471beece1babc095d92-768x446.png 768w\" sizes=\"auto, (max-width: 1024px) 100vw, 1024px\" \/><\/p>\n<h4>2\u3001\u672c\u5730\u7248\u672c\u548c\u57fa\u4e8e\u4e91\u7684\u7248\u672c<\/h4>\n<p>\u5bf9\u4e8e\u90a3\u4e9b\u66f4\u559c\u6b22\u672c\u5730\u90e8\u7f72\u548c\u81ea\u4e3b\u63a7\u5236\u7684\u7528\u6237\uff0c\u53ef\u4ee5\u4ece GitHub \u5b58\u50a8\u5e93\u4e2d\u83b7\u53d6\u8bbf\u95ee\u6743\u9650\uff0c\u5e76\u5c06 Arthur Bench \u90e8\u7f72\u5230\u81ea\u5df1\u7684\u672c\u5730\u73af\u5883\u4e2d\u3002\u8fd9\u6837\uff0c\u5927\u5bb6\u53ef\u4ee5\u5b8c\u5168\u638c\u63e1\u548c\u63a7\u5236 Arthur Bench \u7684\u8fd0\u884c\uff0c\u5e76\u6839\u636e\u81ea\u5df1\u7684\u9700\u6c42\u8fdb\u884c\u5b9a\u5236\u548c\u914d\u7f6e\u3002<\/p>\n<p>\u53e6\u4e00\u65b9\u9762\uff0c\u5bf9\u4e8e\u90a3\u4e9b\u66f4\u503e\u5411\u4e8e\u4fbf\u6377\u548c\u7075\u6d3b\u6027\u7684\u7528\u6237\uff0c\u4e5f\u63d0\u4f9b\u4e86\u57fa\u4e8e\u4e91\u7684 SaaS \u4ea7\u54c1\u3002\u5927\u5bb6\u53ef\u4ee5\u9009\u62e9\u6ce8\u518c\uff0c\u901a\u8fc7\u4e91\u7aef\u8bbf\u95ee\u548c\u4f7f\u7528 Arthur Bench\u3002\u8fd9\u79cd\u65b9\u5f0f\u65e0\u9700\u7e41\u7410\u7684\u672c\u5730\u5b89\u88c5\u548c\u914d\u7f6e\uff0c\u800c\u662f\u80fd\u591f\u7acb\u5373\u4eab\u53d7\u5230\u6240\u63d0\u4f9b\u7684\u529f\u80fd\u548c\u670d\u52a1\u3002<\/p>\n<h4>3\u3001\u5b8c\u5168\u5f00\u6e90<\/h4>\n<p>Arthur Bench \u4f5c\u4e3a\u4e00\u4e2a\u5f00\u6e90\u9879\u76ee\uff0c\u5728\u900f\u660e\u6027\u3001\u53ef\u6269\u5c55\u6027\u548c\u793e\u533a\u534f\u4f5c\u7b49\u65b9\u9762\u5c55\u73b0\u51fa\u5176\u5178\u578b\u7684\u5f00\u6e90\u7279\u70b9\u3002\u8fd9\u79cd\u5f00\u6e90\u6027\u8d28\u4e3a\u7528\u6237\u63d0\u4f9b\u4e86\u4e30\u5bcc\u7684\u4f18\u52bf\u548c\u673a\u4f1a\uff0c\u4f7f\u4ed6\u4eec\u80fd\u591f\u66f4\u6df1\u5165\u5730\u4e86\u89e3\u9879\u76ee\u7684\u5de5\u4f5c\u539f\u7406\uff0c\u5e76\u6839\u636e\u81ea\u8eab\u9700\u6c42\u8fdb\u884c\u5b9a\u5236\u548c\u6269\u5c55\u3002\u540c\u65f6\uff0cArthur Bench \u7684\u5f00\u653e\u6027\u8fd8\u9f13\u52b1\u7528\u6237\u79ef\u6781\u53c2\u4e0e\u793e\u533a\u534f\u4f5c\uff0c\u4e0e\u5176\u4ed6\u7528\u6237\u5171\u540c\u5408\u4f5c\u548c\u53d1\u5c55\u3002\u8fd9\u79cd\u5f00\u653e\u5f0f\u7684\u5408\u4f5c\u6a21\u5f0f\u6709\u52a9\u4e8e\u63a8\u52a8\u9879\u76ee\u7684\u4e0d\u65ad\u53d1\u5c55\u548c\u521b\u65b0\uff0c\u540c\u65f6\uff0c\u4e5f\u4e3a\u7528\u6237\u521b\u9020\u4e86\u66f4\u5927\u7684\u4ef7\u503c\u548c\u673a\u9047\u3002<\/p>\n<p>\u603b\u4e4b\uff0cArthur Bench \u63d0\u4f9b\u4e86\u4e00\u4e2a\u5f00\u653e\u4e14\u7075\u6d3b\u7684\u6846\u67b6\uff0c\u4f7f\u7528\u6237\u80fd\u591f\u81ea\u5b9a\u4e49\u8bc4\u4f30\u6307\u6807\uff0c\u5e76\u4e14\u5df2\u7ecf\u5728\u91d1\u878d\u9886\u57df\u5f97\u5230\u5e7f\u6cdb\u5e94\u7528\u3002\u4e0e Amazon Web Services \u548c Cohere \u7684\u5408\u4f5c\u8fdb\u4e00\u6b65\u63a8\u52a8\u4e86\u8be5\u6846\u67b6\u7684\u53d1\u5c55\uff0c\u9f13\u52b1\u5f00\u53d1\u4eba\u5458\u4e3a Bench \u521b\u5efa\u65b0\u7684\u6307\u6807\uff0c\u4e3a\u8bed\u8a00\u6a21\u578b\u8bc4\u4f30\u9886\u57df\u7684\u8fdb\u6b65\u505a\u51fa\u8d21\u732e\u3002<\/p>\n<p>Reference \uff1a<\/p>\n<ul data-id=\"u738a58b-fYmAuHdQ\">\n<li data-id=\"ld70c578-XNR7jEyE\">[1] https:\/\/github.com\/arthur-ai\/bench<\/li>\n<li data-id=\"ld70c578-dmOOfUXv\">[2] https:\/\/neurohive.io\/en\/news\/arthur-bench-framework-for-evaluating-language-models\/<\/li>\n<\/ul>\n<p>\u6587\u7ae0\u6765\u81ea\uff1a51CTO<\/p>\n<\/div>\n<\/div>\n<div class=\"pvc_clear\"><\/div>\n<p id=\"pvc_stats_18245\" class=\"pvc_stats total_only  \" data-element-id=\"18245\" style=\"\"><i class=\"pvc-stats-icon medium\" aria-hidden=\"true\"><svg xmlns=\"http:\/\/www.w3.org\/2000\/svg\" version=\"1.0\" viewBox=\"0 0 502 315\" preserveAspectRatio=\"xMidYMid meet\"><g transform=\"translate(0,332) scale(0.1,-0.1)\" fill=\"\" stroke=\"none\"><path d=\"M2394 3279 l-29 -30 -3 -207 c-2 -182 0 -211 15 -242 39 -76 157 -76 196 0 15 31 17 60 15 243 l-3 209 -33 29 c-26 23 -41 29 -80 29 -41 0 -53 -5 -78 -31z\"\/><path d=\"M3085 3251 c-45 -19 -58 -50 -96 -229 -47 -217 -49 -260 -13 -295 52 -53 146 -42 177 20 16 31 87 366 87 410 0 70 -86 122 -155 94z\"\/><path d=\"M1751 3234 c-13 -9 -29 -31 -37 -50 -12 -29 -10 -49 21 -204 19 -94 39 -189 45 -210 14 -50 54 -80 110 -80 34 0 48 6 76 34 21 21 34 44 34 59 0 14 -18 113 -40 219 -37 178 -43 195 -70 221 -36 32 -101 37 -139 11z\"\/><path d=\"M1163 3073 c-36 -7 -73 -59 -73 -102 0 -56 133 -378 171 -413 34 -32 83 -37 129 -13 70 36 67 87 -16 290 -86 209 -89 214 -129 231 -35 14 -42 15 -82 7z\"\/><path d=\"M3689 3066 c-15 -9 -33 -30 -42 -48 -48 -103 -147 -355 -147 -375 0 -98 131 -148 192 -74 13 15 57 108 97 206 80 196 84 226 37 273 -30 30 -99 39 -137 18z\"\/><path d=\"M583 2784 c-38 -19 -67 -74 -58 -113 9 -42 211 -354 242 -373 16 -10 45 -18 66 -18 51 0 107 52 107 100 0 39 -1 41 -124 234 -80 126 -108 162 -133 173 -41 17 -61 16 -100 -3z\"\/><path d=\"M4250 2784 c-14 -9 -74 -91 -133 -183 -95 -150 -107 -173 -107 -213 0 -55 33 -94 87 -104 67 -13 90 8 211 198 130 202 137 225 78 284 -27 27 -42 34 -72 34 -22 0 -50 -8 -64 -16z\"\/><path d=\"M2275 2693 c-553 -48 -1095 -270 -1585 -649 -135 -104 -459 -423 -483 -476 -23 -49 -22 -139 2 -186 73 -142 361 -457 571 -626 285 -228 642 -407 990 -497 242 -63 336 -73 660 -74 310 0 370 5 595 52 535 111 1045 392 1455 803 122 121 250 273 275 326 19 41 19 137 0 174 -41 79 -309 363 -465 492 -447 370 -946 591 -1479 653 -113 14 -422 18 -536 8z m395 -428 c171 -34 330 -124 456 -258 112 -119 167 -219 211 -378 27 -96 24 -300 -5 -401 -72 -255 -236 -447 -474 -557 -132 -62 -201 -76 -368 -76 -167 0 -236 14 -368 76 -213 98 -373 271 -451 485 -162 444 86 934 547 1084 153 49 292 57 452 25z m909 -232 c222 -123 408 -262 593 -441 76 -74 138 -139 138 -144 0 -16 -233 -242 -330 -319 -155 -123 -309 -223 -461 -299 l-81 -41 32 46 c18 26 49 83 70 128 143 306 141 649 -6 957 -25 52 -61 116 -79 142 l-34 47 45 -20 c26 -10 76 -36 113 -56z m-2057 25 c-40 -58 -105 -190 -130 -263 -110 -324 -59 -707 132 -981 25 -35 42 -64 37 -64 -19 0 -241 119 -326 174 -188 122 -406 314 -532 468 l-58 71 108 103 c185 178 428 349 672 473 66 33 121 60 123 61 2 0 -10 -19 -26 -42z\"\/><path d=\"M2375 1950 c-198 -44 -350 -190 -395 -379 -18 -76 -8 -221 19 -290 114 -284 457 -406 731 -260 98 52 188 154 231 260 27 69 37 214 19 290 -38 163 -166 304 -326 360 -67 23 -215 33 -279 19z\"\/><\/g><\/svg><\/i> <img loading=\"lazy\" decoding=\"async\" width=\"16\" height=\"16\" alt=\"Loading\" src=\"https:\/\/aif.amtbbs.org\/wp-content\/plugins\/page-views-count\/ajax-loader-2x.gif\" border=0 \/><\/p>\n<div class=\"pvc_clear\"><\/div>\n","protected":false},"excerpt":{"rendered":"<p>\u968f\u7740 LLM \u7684\u5feb\u901f\u53d1\u5c55\u548c\u6539\u8fdb\uff0c\u6211\u4eec\u6b63\u5728\u9762\u5bf9\u65b0\u7684\u6311\u6218\u548c\u673a\u9047\u3002LLM \u7684\u80fd\u529b\u548c\u8868\u73b0\u6c34\u5e73\u4e0d\u65ad\u63d0\u9ad8\uff0c\u8fd9\u4f7f\u5f97\u57fa\u4e8e\u5355\u8bcd\u51fa [&hellip;]<\/p>\n<div class=\"pvc_clear\"><\/div>\n<p id=\"pvc_stats_18245\" class=\"pvc_stats total_only  \" data-element-id=\"18245\" style=\"\"><i class=\"pvc-stats-icon medium\" aria-hidden=\"true\"><svg xmlns=\"http:\/\/www.w3.org\/2000\/svg\" version=\"1.0\" viewBox=\"0 0 502 315\" preserveAspectRatio=\"xMidYMid meet\"><g transform=\"translate(0,332) scale(0.1,-0.1)\" fill=\"\" stroke=\"none\"><path d=\"M2394 3279 l-29 -30 -3 -207 c-2 -182 0 -211 15 -242 39 -76 157 -76 196 0 15 31 17 60 15 243 l-3 209 -33 29 c-26 23 -41 29 -80 29 -41 0 -53 -5 -78 -31z\"\/><path d=\"M3085 3251 c-45 -19 -58 -50 -96 -229 -47 -217 -49 -260 -13 -295 52 -53 146 -42 177 20 16 31 87 366 87 410 0 70 -86 122 -155 94z\"\/><path d=\"M1751 3234 c-13 -9 -29 -31 -37 -50 -12 -29 -10 -49 21 -204 19 -94 39 -189 45 -210 14 -50 54 -80 110 -80 34 0 48 6 76 34 21 21 34 44 34 59 0 14 -18 113 -40 219 -37 178 -43 195 -70 221 -36 32 -101 37 -139 11z\"\/><path d=\"M1163 3073 c-36 -7 -73 -59 -73 -102 0 -56 133 -378 171 -413 34 -32 83 -37 129 -13 70 36 67 87 -16 290 -86 209 -89 214 -129 231 -35 14 -42 15 -82 7z\"\/><path d=\"M3689 3066 c-15 -9 -33 -30 -42 -48 -48 -103 -147 -355 -147 -375 0 -98 131 -148 192 -74 13 15 57 108 97 206 80 196 84 226 37 273 -30 30 -99 39 -137 18z\"\/><path d=\"M583 2784 c-38 -19 -67 -74 -58 -113 9 -42 211 -354 242 -373 16 -10 45 -18 66 -18 51 0 107 52 107 100 0 39 -1 41 -124 234 -80 126 -108 162 -133 173 -41 17 -61 16 -100 -3z\"\/><path d=\"M4250 2784 c-14 -9 -74 -91 -133 -183 -95 -150 -107 -173 -107 -213 0 -55 33 -94 87 -104 67 -13 90 8 211 198 130 202 137 225 78 284 -27 27 -42 34 -72 34 -22 0 -50 -8 -64 -16z\"\/><path d=\"M2275 2693 c-553 -48 -1095 -270 -1585 -649 -135 -104 -459 -423 -483 -476 -23 -49 -22 -139 2 -186 73 -142 361 -457 571 -626 285 -228 642 -407 990 -497 242 -63 336 -73 660 -74 310 0 370 5 595 52 535 111 1045 392 1455 803 122 121 250 273 275 326 19 41 19 137 0 174 -41 79 -309 363 -465 492 -447 370 -946 591 -1479 653 -113 14 -422 18 -536 8z m395 -428 c171 -34 330 -124 456 -258 112 -119 167 -219 211 -378 27 -96 24 -300 -5 -401 -72 -255 -236 -447 -474 -557 -132 -62 -201 -76 -368 -76 -167 0 -236 14 -368 76 -213 98 -373 271 -451 485 -162 444 86 934 547 1084 153 49 292 57 452 25z m909 -232 c222 -123 408 -262 593 -441 76 -74 138 -139 138 -144 0 -16 -233 -242 -330 -319 -155 -123 -309 -223 -461 -299 l-81 -41 32 46 c18 26 49 83 70 128 143 306 141 649 -6 957 -25 52 -61 116 -79 142 l-34 47 45 -20 c26 -10 76 -36 113 -56z m-2057 25 c-40 -58 -105 -190 -130 -263 -110 -324 -59 -707 132 -981 25 -35 42 -64 37 -64 -19 0 -241 119 -326 174 -188 122 -406 314 -532 468 l-58 71 108 103 c185 178 428 349 672 473 66 33 121 60 123 61 2 0 -10 -19 -26 -42z\"\/><path d=\"M2375 1950 c-198 -44 -350 -190 -395 -379 -18 -76 -8 -221 19 -290 114 -284 457 -406 731 -260 98 52 188 154 231 260 27 69 37 214 19 290 -38 163 -166 304 -326 360 -67 23 -215 33 -279 19z\"\/><\/g><\/svg><\/i> <img loading=\"lazy\" decoding=\"async\" width=\"16\" height=\"16\" alt=\"Loading\" src=\"https:\/\/aif.amtbbs.org\/wp-content\/plugins\/page-views-count\/ajax-loader-2x.gif\" border=0 \/><\/p>\n<div class=\"pvc_clear\"><\/div>\n","protected":false},"author":56,"featured_media":18246,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[3,23,80],"tags":[124,986,113],"class_list":["post-18245","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ai","category-23","category-80","tag-ai","tag-llm","tag-113"],"_links":{"self":[{"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/posts\/18245","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/users\/56"}],"replies":[{"embeddable":true,"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/comments?post=18245"}],"version-history":[{"count":1,"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/posts\/18245\/revisions"}],"predecessor-version":[{"id":18250,"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/posts\/18245\/revisions\/18250"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/media\/18246"}],"wp:attachment":[{"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/media?parent=18245"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/categories?post=18245"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/tags?post=18245"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}