{"id":2970,"date":"2023-08-13T18:19:48","date_gmt":"2023-08-13T10:19:48","guid":{"rendered":""},"modified":"2023-08-13T18:19:48","modified_gmt":"2023-08-13T10:19:48","slug":"","status":"publish","type":"post","link":"http:\/\/www.idea2003.com\/dongtai\/2970.html","title":{"rendered":"GPT-4\u63a8\u7406\u592a\u79bb\u8c31\uff01\u5927\u5b66\u6570\u7406\u5316\u603b\u5206\u6ca1\u8fc7\u534a\uff0c21\u7c7b\u63a8\u7406\u9898\u5168\u7ffb\u8f66\uff0c\u9a6c\u5e93\u65af\uff1aAGI\u592a\u9065\u8fdc"},"content":{"rendered":"

\u7f16\u8005\u6309\uff1a\u672c\u6587\u6765\u81ea\u5fae\u4fe1\u516c\u4f17\u53f7 \u65b0\u667a\u5143\uff08ID\uff1aAI_era\uff09\uff0c\u7f16\u8f91\uff1a\u7f16\u8f91\u90e8\uff0c\u5fae\u65b0\u521b\u60f3\u7ecf\u6388\u6743\u8f6c\u8f7d\u3002<\/p>\n

GPT-4\u6839\u672c\u4e0d\u4f1a\u63a8\u7406\uff01<\/p>\n

\u8fd1\u6765\uff0c\u6709\u4e24\u7bc7\u7814\u7a76\u79f0\uff0cGPT-4\u5728\u63a8\u7406\u65b9\u9762\u8868\u73b0\u4e0d\u5c3d\u4eba\u610f\u3002<\/p>\n

\u6765\u81eaMIT\u7684\u6821\u53cbKonstantine Arkoudas\uff0c\u572821\u79cd\u4e0d\u540c\u7c7b\u578b\u63a8\u7406\u96c6\u4e2d\uff0c\u5bf9GPT-4\u8fdb\u884c\u4e86\u8bc4\u4f30\u3002<\/p>\n

\u7136\u540e\uff0c\u5bf9GPT-4\u5728\u8fd9\u4e9b\u95ee\u9898\u4e0a\u7684\u8868\u73b0\u8fdb\u884c\u4e86\u8be6\u7ec6\u7684\u5b9a\u6027\u5206\u6790\u3002<\/p>\n

\u7814\u7a76\u53d1\u73b0\uff0cGPT-4\u5076\u5c14\u4f1a\u5c55\u73b0\u51fa\u300c\u6700\u5f3a\u5927\u8111\u300d\u7684\u5929\u8d4b\uff0c\u4f46\u76ee\u524d\u6765\u770b\uff0cGPT-4\u5b8c\u5168\u4e0d\u5177\u5907\u63a8\u7406\u80fd\u529b\u3002<\/p>\n

\"\u56fe\u7247\"<\/p>\n

\u8bba\u6587\u5730\u5740\uff1ahttps:\/\/www.preprints.org\/manuscript\/202308.0148\/v2<\/p>\n

\u7814\u7a76\u4e00\u51fa\uff0c\u5f15\u6765\u4f17\u591a\u7f51\u53cb\u56f4\u89c2\u3002<\/p>\n

\u9a6c\u5e93\u65af\u8868\u793a\uff0c\u300c\u5982\u679c\u8fd9\u662f\u771f\u7684\u2014\u2014\u6b63\u5982\u6211\u65e9\u5c31\u8bf4\u8fc7\u7684\u90a3\u6837\u2014\u2014\u6211\u4eec\u79bbAGI\u8fd8\u5dee\u5f97\u8fdc\u5462\u3002\u6211\u4eec\u53ef\u80fd\u9700\u8981\u8fdb\u884c\u5927\u91cf\u7684\u91cd\u65b0\u6821\u51c6\uff1a\u6ca1\u6709\u63a8\u7406\u5c31\u4e0d\u53ef\u80fd\u6709 AGI\u300d\u3002<\/p>\n

\"\u56fe\u7247\"<\/p>\n

\u800c\u53e6\u4e00\u7bc7\u6765\u81eaUCLA\u548c\u534e\u76db\u987f\u5927\u5b66\u7684\u7814\u7a76\u4e5f\u53d1\u73b0\uff0cGPT-4\uff0c\u4ee5\u53caGPT-3.5\u5728\u5927\u5b66\u7684\u6570\u5b66\u3001\u7269\u7406\u3001\u5316\u5b66\u4efb\u52a1\u7684\u63a8\u7406\u4e0a\uff0c\u8868\u73b0\u4e0d\u4f73\u3002<\/p>\n

\"\u56fe\u7247\"<\/p>\n

\u8bba\u6587\u5730\u5740\uff1ahttps:\/\/arxiv.org\/pdf\/2307.10635.pdf<\/p>\n

\u7814\u7a76\u4eba\u5458\u5f15\u5165\u4e86\u4e00\u4e2a\u5927\u5b66\u79d1\u5b66\u95ee\u9898\u89e3\u51b3\u57fa\u7840SCIBENCH\uff0c\u5176\u4e2d\u5305\u542b2\u4e2a\u6570\u636e\u96c6\uff1a\u5f00\u653e\u6570\u636e\u96c6\uff0c\u4ee5\u53ca\u5c01\u95ed\u6570\u636e\u96c6\u3002<\/p>\n

\u901a\u8fc7\u5bf9GPT-4\u548cGPT-3.5\u91c7\u7528\u4e0d\u540c\u63d0\u793a\u7b56\u7565\u8fdb\u884c\u6df1\u5165\u7814\u7a76\uff0c\u7ed3\u679c\u663e\u793a\uff0cGPT-4\u6210\u7ee9\u5e73\u5747\u603b\u5206\u4ec5\u4e3a35.8%\u3002<\/p>\n

\u8fd9\u9879\u7814\u7a76\u540c\u6837\u518d\u6b21\u5f15\u8d77\u9a6c\u5e93\u65af\u7684\u5173\u6ce8\uff1a<\/p>\n

\n

\u5173\u4e8e\u6570\u5b66\u3001\u5316\u5b66\u548c\u7269\u7406\u63a8\u7406\u7684\u7cfb\u7edf\u8c03\u67e5\uff0c\u7ed3\u679c\u663e\u793a\uff0c\u76ee\u524d\u7684LLM\u65e0\u6cd5\u63d0\u4f9b\u4ee4\u4eba\u6ee1\u610f\u7684\u6027\u80fd……\u6ca1\u6709\u4e00\u79cd\u63d0\u793a\u7b56\u7565\u660e\u663e\u4f18\u4e8e\u5176\u4ed6\u7b56\u7565\u3002<\/p>\n<\/blockquote>\n

\"\u56fe\u7247\"<\/p>\n

\u4e0b\u9762\u6211\u4eec\u5c31\u6765\u5177\u4f53\u770b\u770b\uff0cGPT-4\u5982\u4f55\u572821\u4e2a\u95ee\u9898\u96c6\uff0c\u6570\u5b66\u3001\u7269\u7406\u3001\u5316\u5b66\u4e0a\u63a8\u7406\u60e8\u8d25\u7684\u3002<\/p>\n

21\u4e2a\u95ee\u9898\u96c6\uff0cGPT-4\u5168\u7ffb\u8f66<\/h2>\n

\u4e0d\u8fc7\uff0c\u5728\u770bGPT-4\u56de\u7b54\u95ee\u9898\u4e4b\u524d\uff0c\u4f5c\u8005\u7ed9\u51fa\u4e86\u4e00\u4e2a\u6ce8\u610f\u4e8b\u9879\uff1a<\/p>\n

GPT-4\u662f\u4e00\u4e2a\u975e\u786e\u5b9a\u6027\u7cfb\u7edf\uff0c\u5373\u4f7f\u53c2\u6570\u8bbe\u7f6e\u76f8\u540c\uff0c\u5728\u4e0d\u540c\u7684\u8fd0\u884c\u4e2d\u4e5f\u53ef\u80fd\u4ea7\u751f\u4e0d\u540c\u7684\u7b54\u6848\u3002<\/p>\n

\u800c\u4ee5\u4e0b\u7684\u6d4b\u8bd5\u4ea4\u6d41\u90fd\u662f\u9010\u5b57\u8bb0\u5f55\u7684\uff0c\u6839\u636e\u4f5c\u8005\u7684\u7ecf\u9a8c\uff0c\u6587\u4e2d\u8ba8\u8bba\u7684GPT-4\u51fa\u9519\u7684\u5730\u65b9\u5f80\u5f80\u5177\u6709\u9c81\u68d2\u6027\u3002<\/p>\n

1. \u7b80\u5355\u7b97\u672f<\/strong><\/p>\n

\u80fd\u591f\u8fdb\u884c\u57fa\u672c\u8fd0\u7b97\uff0c\u662f\u63a8\u7406\u7684\u5fc5\u8981\u6761\u4ef6\u3002<\/p>\n

\u4f46\u662f\uff0cGPT-4\u4ecd\u7136\u65e0\u6cd5\u53ef\u9760\u5730\u6267\u884c\u52a0\u6cd5\u3001\u4e58\u6cd5\u7b49\u57fa\u672c\u7b97\u672f\u8fd0\u7b97\u3002<\/p>\n

\u6bd4\u5982\uff0c\u8ba9GPT-4\u57281381\u548c1453\u4e4b\u95f4\u968f\u673a\u9009\u62e9\u4e24\u4e2a\u6570\u5b57\u76f8\u4e58\uff0c\u5e76\u7ed9\u51fa\u7ed3\u679c\u3002<\/p>\n

GPT-4\u9009\u62e9\u4e861405\uff0c\u4ee5\u53ca1421\uff0c\u4f46\u662f\u6700\u540e\u7ed9\u51fa\u7684\u7ed3\u679c\u663e\u7136\u662f\u9519\u7684\u3002\u56e0\u4e3a1405\u00d71421=1996505\u3002<\/p>\n

\"\u56fe\u7247\"<\/p>\n

2. \u7b80\u5355\u8ba1\u6570<\/strong><\/p>\n

\u867d\u7136\u5177\u4f53\u8ba1\u6570\u5e76\u4e0d\u4e00\u5b9a\u662f\u4e00\u79cd\u63a8\u7406\u6d3b\u52a8 \uff0c\u4f46\u5b83\u80af\u5b9a\u662f\u4efb\u4f55\u5177\u6709\u4e00\u822c\u80fd\u529b\u63a8\u7406\u7cfb\u7edf\u7684\u5fc5\u5907\u6761\u4ef6\u3002<\/p>\n

\u5728\u8fd9\u91cc\uff0c\u7ed9GPT-4\u4e00\u4e2a\u547d\u9898\u53d8\u91cf\uff0c\u5e76\u5728\u5b83\u524d\u9762\u52a0\u4e0a27\u4e2a\u5426\u5b9a\u7b26\u53f7\uff0c\u8981\u6c42\u5b83\u8ba1\u7b97\u5426\u5b9a\u7b26\u53f7\u7684\u4e2a\u6570\u3002<\/p>\n

\u5bf9\u4e8e\u6211\u4eec\u6765\u8bb2\uff0c\u8fd9\u7b80\u76f4\u8f7b\u800c\u6613\u4e3e\uff0c\u5c24\u5176\u662f\u5426\u5b9a\u7b26\u53f7\u662f\u95f4\u96945\u4e2a\u5199\u6210\u7684\uff0c\u5e76\u4e14\u67095\u7ec4\uff0c\u6700\u540e\u4e00\u5bf9\u5426\u5b9a\u7b26\u53f7\u7d27\u968f\u5176\u540e\u3002<\/p>\n

\u7136\u800c\uff0cGPT-4\u5374\u7ed9\u51fa\u4e86\u300c28\u4e2a\u300d\u7b54\u6848\u3002<\/p>\n

\"\u56fe\u7247\"<\/p>\n

3. \uff08\u533b\u5b66\uff09\u5e38\u8bc6<\/strong><\/p>\n

\u5f53\u524d\uff0c\u6211\u4eec\u53ef\u4ee5\u5c06\u5e38\u8bc6\u6027\u8bba\u8bc1\u89c6\u4e3a\uff0c\u4ece\u7ed9\u5b9a\u4fe1\u606f\u52a0\u4e0a\u672a\u8bf4\u660e\u7684\u6761\u4ef6\uff08\u9ed8\u8ba4\u7684\u3001\u666e\u904d\u63a5\u53d7\u7684\u80cc\u666f\u77e5\u8bc6\uff09\u4e2d\u5f97\u51fa\u7684\u7b80\u5355\u63a8\u7406\u3002<\/p>\n

\u5728\u8fd9\u79cd\u7279\u6b8a\u60c5\u51b5\u4e0b\uff0c\u5e38\u8bc6\u6027\u77e5\u8bc6\u5c31\u662f\u300c\u4eba\u5728\u6b7b\u524d\u662f\u6d3b\u7740\u7684\uff0c\u6b7b\u540e\u5c31\u4e0d\u4f1a\u518d\u6d3b\u7740\u300d\u8fd9\u6837\u7684\u547d\u9898\u3002<\/p>\n

\u6bd4\u5982\uff0c\u5f53\u4f60\u95eeGPT-4\uff1aMable\u4e0a\u53489\u70b9\u7684\u5fc3\u7387\u4e3a75 bpm\uff0c\u4e0b\u53487\u70b9\u7684\u8840\u538b\u4e3a120\/80\u3002\u5979\u4e8e\u665a\u4e0a11\u70b9\u6b7b\u4ea1\u3002\u5979\u4e2d\u5348\u8fd8\u6d3b\u7740\u5417\uff1f<\/p>\n

\"\u56fe\u7247\"<\/p>\n

GPT-4\u7adf\u56de\u7b54\uff1a\u6839\u636e\u6240\u63d0\u4f9b\u7684\u4fe1\u606f\uff0c\u65e0\u6cd5\u786e\u5b9aMable\u4e2d\u5348\u662f\u5426\u8fd8\u6d3b\u7740\u3002<\/p>\n

\u4f46\u660e\u663e\u6839\u636e\u7ed9\u5b9a\u7684\u4fe1\u606f\uff0c\u5e38\u8bc6\u6027\u63a8\u65ad\uff08\u4e0d\u7528\u60f3\uff09\u76f4\u63a5\u5f97\u51fa\u7ed3\u8bba\u4e86\u3002<\/p>\n

4. \u521d\u7ea7\u903b\u8f91<\/strong><\/p>\n

\u5982\u679cP(x)\u5305\u542bQ(x)\uff0c\u800cQ(a)\u4e0d\u6210\u7acb\uff0c\u90a3\u4e48\u6211\u4eec\u5c31\u53ef\u4ee5\u6839\u636e\u6a21\u578b\u63a8\u8bba\u51faP(a)\u4e5f\u4e0d\u6210\u7acb\uff08\u56e0\u4e3a\u5982\u679cP(a)\u6210\u7acb\uff0c\u90a3\u4e48Q(a)\u4e5f\u4f1a\u6210\u7acb\uff09\u3002<\/p>\n

\u8fd9\u662f\u4e00\u4e2a\u6700\u57fa\u672c\u7684\u540c\u4e49\u53cd\u590d\uff0c\u4f46GPT-4\u5374\u5b8c\u5168\u63d0\u51fa\u4e00\u4e2a\u53cd\u6a21\u578b\uff1a<\/p>\n

\"\u56fe\u7247\"<\/p>\n

\u503c\u5f97\u6ce8\u610f\u7684\u662f\uff0cGPT-4\u8ba4\u8bc6\u5230\uff0cP(x)\u5b9e\u9645\u4e0a\u5e76\u4e0d\u5305\u542bQ(x)\uff0c\u5e76\u63d0\u51fa\u4e86x\u6709\u53ef\u80fd\u662f\u8d1f\u6570\u5076\u6570\uff0c\u300c\u4e0d\u6392\u9664\u5b58\u5728\u5176\u4ed6\u7ed9\u5b9a\u6761\u4ef6\u7684\u6a21\u578b\u300d\u3002<\/p>\n

\u5176\u5b9e\u4e0d\u7136\uff0c\u4e00\u4e2a\u53cd\u6a21\u578b\uff08countermodel\uff09\u5fc5\u987b\u6ee1\u8db3\u6240\u6709\u7ed9\u5b9a\u7684\u6761\u4ef6\uff0c\u540c\u65f6\u8bc1\u4f2a\u7ed3\u8bba\u3002<\/p>\n

\u6b64\u5916\uff0c\u4ec5\u4ec5\u51e0\u53e5\u8bdd\u4e4b\u540e\uff0c GPT-4\u5c31\u58f0\u79f0P(x)\u5728\u7ed9\u5b9a\u7684\u89e3\u91ca\u4e0b\u786e\u5b9e\u8574\u542bQ(x)\uff0c\u8fd9\u4e0e\u5b83\u81ea\u5df1\u4e4b\u524d\u7684\u8bf4\u6cd5\u76f8\u77db\u76fe\u3002<\/p>\n

\"\u56fe\u7247\"<\/p>\n

\u8bf4\u660e\uff0c GPT-4\u8fd8\u4f1a\u51fa\u73b0\u5185\u90e8\u4e0d\u4e00\u81f4\u7684\u95ee\u9898\u3002<\/p>\n

5. \u7b80\u5355\u91cf\u8bcd\u8bed\u4e49<\/strong><\/p>\n

\u8bf7\u770b\u4e0b\u9762\u4e09\u4e2a\u53e5\u5b50\uff1a<\/p>\n

1. [forall x . P(x) ==> Q(x)]<\/p>\n

2. [exists x . P(x)]<\/p>\n

3. [exists x . \u223c Q(x)]<\/p>\n

\u8bf7\u8bc1\u4f2a\u6216\u8bc1\u660e\u4ee5\u4e0b\u4e3b\u5f20\uff1a\u8fd9\u4e09\u4e2a\u53e5\u5b50\u662f\u5171\u540c\u53ef\u6ee1\u8db3\u7684\u3002<\/p>\n

\"\u56fe\u7247\"<\/p>\n

\u663e\u7136\uff0c\u8fd9\u4e09\u4e2a\u53e5\u5b50\u90fd\u662f\u5171\u540c\u53ef\u6ee1\u8db3\u7684\uff0c\u4e00\u4e2a\u7b80\u5355\u7684\u6a21\u578b\u662f\u5177\u6709P(a1)\u3001Q(a1)\u3001\u00acP(a2) \u548c \u00acQ(a2)\u7684\u57df{a1, a2}\uff0c\u7136\u800cGPT-4\u5f97\u51fa\u7684\u7ed3\u8bba\u786e\u4e0e\u4e4b\u76f8\u53cd\u3002<\/p>\n

6. \u7b80\u5355\u56fe\u7740\u8272<\/strong><\/p>\n

\u9996\u5148\u8003\u8651\u4e00\u4e2a\u6ca1\u6709\u89e3\u51b3\u65b9\u6848\u7684\u56fe\u7740\u8272\u95ee\u9898\u3002<\/p>\n

\u4e0d\u96be\u53d1\u73b0\uff0c\u5bf9\u4e8e\u8fd9\u4e2a\u95ee\u9898\u4e2d\u63cf\u8ff0\u7684\u56fe\u5f62\uff0c\u4e24\u79cd\u989c\u8272\u662f\u4e0d\u8db3\u4ee5\u6ee1\u8db3\u95ee\u9898\u4e2d\u63cf\u8ff0\u7684\u56fe\uff08\u4f8b\u5982\uff0c\u9876\u70b90\u30012\u548c4\u5f62\u6210\u4e86\u4e00\u4e2a\u7c07\uff0c\u56e0\u6b64\u81f3\u5c11\u9700\u89813\u79cd\u989c\u8272\uff09\u3002<\/p>\n

\"\u56fe\u7247\"<\/p>\n

\u5728\u8fd9\u4e2a\u7b80\u77ed\u7684\u8f93\u51fa\u4e2d\uff0c\u51fa\u73b0\u5927\u91cf\u60ca\u5413\u4e0b\u5df4\u7684\u9519\u8bef\u3002<\/p>\n

GPT-4\u4e00\u5f00\u59cb\u5c31\u8c0e\u79f0\u56fe\u5f62\u662f\u5b8c\u5168\u7684\uff08\u663e\u7136\u4e0d\u662f\uff0c\u4f8b\u5982\u9876\u70b92\u548c3\u4e4b\u95f4\u6ca1\u6709\u8fb9\uff09\u3002<\/p>\n

\u6b64\u5916\uff0c\u663e\u800c\u6613\u89c1\u7684\u662f\uff0c\u5982\u679c\u56fe\u5f62\u771f\u662f\u5b8c\u5168\u7684\uff0c\u90a3\u4e48\u5c31\u4e0d\u53ef\u80fd\u75282\u79cd\u989c\u8272\u6765\u7740\u8272\uff0c\u56e0\u4e3a\u4e00\u4e2a\u67096\u4e2a\u9876\u70b9\u7684\u5b8c\u5168\u56fe\u5f62\u81f3\u5c11\u9700\u89816\u79cd\u989c\u8272\u3002<\/p>\n

\u6362\u53e5\u8bdd\u8bf4\uff0cGPT-4\u7684\u8bf4\u6cd5\u4e0d\u4ec5\u662f\u9519\u8bef\u7684\uff0c\u800c\u4e14\u662f\u524d\u540e\u77db\u76fe\u7684\uff1a\u4e00\u4f1a\u513f\u544a\u8bc9\u6211\u4eec\uff08\u9519\u8bef\uff09\u8fd96\u9876\u70b9\u56fe\u5f62\u662f\u5b8c\u5168\u7684\uff0c\u8fd9\u610f\u5473\u7740\u4e0d\u53ef\u80fd\u75282\u79cd\u989c\u8272\u7ed9\u5b83\u7740\u8272\uff0c\u4e00\u4f1a\u513f\u53c8\u63d0\u4f9b\u4e86\u4e00\u79cd\u53cc\u8272\u300c\u89e3\u51b3\u65b9\u6848\u300d\u3002<\/p>\n

\u503c\u5f97\u6ce8\u610f\u7684\u662f\uff0cGPT-4\u4e4b\u6240\u4ee5\u8868\u73b0\u5982\u6b64\u7cdf\u7cd5\uff0c\u5e76\u4e0d\u662f\u56e0\u4e3a\u5b83\u6ca1\u6709\u638c\u63e1\u8db3\u591f\u7684\u56fe\u5f62\u77e5\u8bc6\u6216\u6570\u636e\u3002<\/p>\n

\u5f53\u7814\u7a76\u4eba\u5458\u8981\u6c42GPT-4\u5bf9\u300c\u5b8c\u5168\u56fe\u300d\u7684\u4e86\u89e3\u65f6\uff0c\u5b83\u6ed4\u6ed4\u4e0d\u7edd\u5730\u8bf4\u51fa\u4e86\u300c\u5b8c\u5168\u56fe\u300d\u7684\u6b63\u786e\u5b9a\u4e49\uff0c\u4ee5\u53ca\u4e00\u957f\u4e32\u5173\u4e8eK_n\uff08\u6709n\u4e2a\u9876\u70b9\u7684\u5b8c\u5168\u56fe\uff09\u7684\u7ed3\u679c\u3002<\/p>\n

\u663e\u7136\uff0cGPT-4 \u5df2\u7ecf\u8bb0\u4f4f\u4e86\u6240\u6709\u8fd9\u4e9b\u4fe1\u606f\uff0c\u4f46\u5374\u65e0\u6cd5\u5728\u65b0\u6761\u4ef6\u4e2d\u5e94\u7528\u3002<\/p>\n

7. \u5b50\u96c6\u548c<\/strong><\/p>\n

S = {2, 8, 6, 32, 22, 44, 28, 12, 18, 10, 14}\u3002\u90a3\u4e48S\u6709\u591a\u5c11\u4e2a\u5b50\u96c6\u7684\u603b\u548c\u662f37\uff1f<\/p>\n

\u8fd9\u4e2a\u95ee\u9898\u4e2d\uff0cS\u7684\u5b50\u96c6\u90fd\u662f\u5076\u6570\uff0c\u800c\u5076\u6570\u4e4b\u548c\u4e0d\u53ef\u80fd\u662f\u5947\u6570\uff0c\u56e0\u6b64\u7b54\u6848\u4e3a0\u3002<\/p>\n

\u7136\u800c\uff0cGPT-4\u6ca1\u6709\u505c\u4e0b\u6765\u8003\u8651S\u5305\u542b\u7684\u5185\u5bb9\uff0c\u800c\u662f\u53cd\u5c04\u6027\u5730\u751f\u6210\u5b83\u8ba4\u4e3a\u5bf9\u8fd9\u4e2a\u95ee\u9898\u5408\u9002\u7684\u7b54\u6848\uff0c\u7136\u540e\u7ee7\u7eed\u300c\u5e7b\u5316\u300d\u51fa\u4e00\u4e2a\u7b54\u6848\u300c4\u300d\u3002<\/p>\n

\"\u56fe\u7247\"<\/p>\n

8. \u521d\u7ea7\u79bb\u6563\u6570\u5b66<\/strong><\/p>\n

\u544a\u8bc9GPT-4 A \u00d7 B\u4ee3\u8868\u96c6\u5408A\u548cB\u7684\u7b1b\u5361\u5c14\u79ef\u3001\u4eceA\u5230B\u7684\u5173\u7cfbR\u662fA \u00d7 B\u7684\u5b50\u96c6\uff0c\u4ee5\u53ca&\u4ee3\u8868\u96c6\u5408\u4ea4\u96c6\u4e4b\u540e\u8981\u6c42\u5b83\u8bc1\u660e\u6216\u8bc1\u4f2a\uff1a<\/p>\n

\"\u56fe\u7247\"<\/p>\n

\u5176\u4e2dR1\u548cR2\u662f\u4eceA\u5230B\u7684\u4e8c\u5143\u5173\u7cfb\uff0cdom(R)\u8868\u793a\u4e8c\u5143\u5173\u7cfbR\u7684\u57df\u3002<\/p>\n

\u9700\u8981\u5b50\u96c6\u5173\u7cfb\u5728(2)\u7684\u4e24\u4e2a\u65b9\u5411\u4e0a\u90fd\u6210\u7acb\uff0c\u4f46\u5b83\u53ea\u5728\u4ece\u5de6\u5230\u53f3\u7684\u65b9\u5411\u4e0a\u6210\u7acb\u3002\u53e6\u4e00\u4e2a\u65b9\u5411\u7684\u53cd\u4f8b\u5f88\u5bb9\u6613\u627e\u5230\uff08\u4f8b\u5982\uff0c\u53d6A = {(1, 2)} \u548c B = {(1,3)}\uff09\u3002<\/p>\n

\u7136\u800c\uff0cGPT-4\u5374\u63a8\u65ad\u8fd9\u662f\u6210\u7acb\u7684\uff0c\u663e\u7136\u4e0d\u6b63\u786e\u3002<\/p>\n

\"\u56fe\u7247\"<\/p>\n

9. \u7b80\u5355\u5b89\u6392\u8ba1\u5212<\/strong><\/p>\n

\u5728\u65f6\u95f4\u5b89\u6392\u95ee\u9898\u4e0a\uff0cGPT-4\u540c\u6837\u51fa\u9519\u4e86\u3002<\/p>\n

\"\u56fe\u7247\"<\/p>\n

<\/p>\n

10. \u7f57\u7d20\u6096\u8bba<\/strong><\/p>\n

\u7f57\u7d20\u7406\u53d1\u5e08\u6096\u8bba\u662f\u6307\uff0c\u5b58\u5728\u4e00\u4e2a\u7406\u53d1\u5e08b\uff0c\u4ed6\u4e3a\u4e14\u4ec5\u4e3a\u90a3\u4e9b\u4e0d\u7ed9\u81ea\u5df1\u522e\u80e1\u5b50\u7684\u4eba\u522e\u80e1\u5b50\u3002<\/p>\n

\u8fd9\u53e5\u8bdd\u7684\u5426\u5b9a\u662f\u4e00\u4e2a\u540c\u4e49\u53cd\u590d\uff0c\u5f88\u5bb9\u6613\u7528\u4e00\u9636\u903b\u8f91\u63a8\u5bfc\u51fa\u6765\u3002<\/p>\n

\u5982\u679c\u6211\u4eec\u628aR(a,b)\u7406\u89e3\u4e3aa\u88abb\u522e\u80e1\u5b50\uff0c\u90a3\u4e48\u6211\u4eec\u5c31\u53ef\u4ee5\u63d0\u51fa\u8fd9\u4e2a\u540c\u4e49\u53cd\u590d\uff0c\u5e76\u8981\u6c42GPT-4\u8bc1\u660e\u6216\u53cd\u8bc1\u5b83\uff0c\u5982\u4e0b\u9762prompt\u6240\u793a\uff1a<\/p>\n

\u5982\u679c\u5b58\u5728\u8fd9\u6837\u4e00\u4e2a\u7406\u53d1\u5e08x\uff0c\u90a3\u4e48\u5bf9\u4e8e\u6240\u6709y\uff0c\u6211\u4eec\u5c06\u6709R(y,x) <==> \u223c R(y,y)\uff0c\u56e0\u6b64\u7528x\u4ee3\u66ffy\u5c06\u5f97\u5230R(x\uff0cx) <==> \u223c R(x,x)\uff0c\u8fd9\u662f\u77db\u76fe\u7684\u3002<\/p>\n

GPT-4\u5bf9\u6240\u7ed9\u53e5\u5b50\u7684\u7ed3\u6784\u548c\u9700\u8981\u505a\u7684\u4e8b\u60c5\u7684\u7406\u89e3\u65e0\u53ef\u6311\u5254\u3002\u7136\u800c\uff0c\u968f\u540e\u7684\u6848\u4f8b\u5206\u6790\u5374\u7cca\u91cc\u7cca\u6d82\u3002<\/p>\n

\"\u56fe\u7247\"<\/p>\n

11. \u79ef\u6728\u4e16\u754c<\/strong><\/p>\n

\u8fd9\u662f\u4e00\u4e2a\u7b80\u5355\u7684\u63a8\u7406\u4efb\u52a1\uff0c\u9700\u8981\u5bf9\u5012\u6570\u7b2c\u4e09\u4e2a\u79ef\u6728B3\u8fdb\u884c\u6848\u4f8b\u5206\u6790\u3002<\/p>\n

\u9996\u5148\uff0cB3\u8981\u4e48\u662f\u7eff\u8272\u7684\uff0c\u8981\u4e48\u4e0d\u662f\u3002<\/p>\n

\u5982\u679c\u662f\u7eff\u8272\u7684\uff0c\u90a3\u4e48B3\u5c31\u5728\u975e\u7eff\u8272\u79ef\u6728B4\u7684\u4e0a\u9762\uff0c\u6240\u4ee5\u7ed3\u8bba\u6210\u7acb\u3002<\/p>\n

\u5982\u679c\u4e0d\u662f\uff0c\u90a3\u4e48\u4ece\u4e0a\u6570\u7684\u7b2c\u4e8c\u4e2a\u7eff\u8272\u79ef\u6728B2\uff0c\u5c31\u5728\u975e\u7eff\u8272\u79ef\u6728B3\u4e0a\u9762\uff0c\u56e0\u6b64\u7ed3\u8bba\u4ecd\u7136\u6210\u7acb\u3002<\/p>\n

\u7136\u800c\uff0c\u7ed3\u679c\u663e\u793a\uff0cGPT-4\u7684\u8868\u73b0\u5e76\u4e0d\u7406\u60f3\u3002<\/p>\n

\"\u56fe\u7247\"<\/p>\n

\n

\u6709\u4e94\u4e2a\u79ef\u6728\u4ece\u4e0a\u5f80\u4e0b\u5806\u53e0\uff1a <\/p>\n

1. \u4ece\u4e0a\u5f80\u4e0b\u6570\u7b2c\u4e8c\u4e2a\u79ef\u6728\u662f\u7eff\u8272\u7684<\/p>\n

2. \u4ece\u4e0a\u5f80\u4e0b\u6570\u7b2c\u56db\u4e2a\u79ef\u6728\u4e0d\u662f\u7eff\u8272\u7684 <\/p>\n

\u5728\u8fd9\u4e9b\u6761\u4ef6\u6210\u7acb\u7684\u60c5\u51b5\u4e0b\uff0c\u8bc1\u4f2a\u6216\u8bc1\u660e\u4ee5\u4e0b\u7ed3\u8bba\uff1a\u5728\u4e00\u4e2a\u975e\u7eff\u8272\u79ef\u6728\u7684\u6b63\u4e0a\u65b9\uff0c\u6709\u4e00\u4e2a\u7eff\u8272\u79ef\u6728\u3002<\/p>\n<\/blockquote>\n

\u9996\u5148\u5b83\u5728\u8bc1\u660e\u731c\u60f3\u65f6\uff0c\u5c31\u5df2\u7ecf\u5f04\u9519\u4e86\u8bc1\u660e\u7684\u7b56\u7565\u2014\u2014PT-4\u5047\u5b9a\u4e86\u4e24\u79cd\u7279\u6b8a\u60c5\u51b5\u6765\u8fdb\u884c\u63a8\u7406\u3002<\/p>\n

\u6b64\u5916\uff0cGPT-4\u5728\u81ea\u5df1\u7684\u63a8\u7406\u4e2d\u5df2\u7ecf\u5f97\u51fa\u4e86\u7ed3\u8bba\uff08\u867d\u7136\u662f\u9519\u7684\uff09\uff0c\u4f46\u5728\u56de\u7b54\u65f6\u4ecd\u7136\u544a\u8bc9\u7528\u6237\u95ee\u9898\u6ca1\u6709\u88ab\u89e3\u51b3\u3002\u800c\u8fd9\u4f53\u73b0\u7684\u4fbf\u662f\u6a21\u578b\u7684\u5185\u90e8\u4e0d\u4e00\u81f4\u6027\u95ee\u9898\u3002<\/p>\n

\"\u56fe\u7247\"<\/p>\n

12. \u7a7a\u95f4\u63a8\u7406<\/strong><\/p>\n

\u8fd9\u91cc\u4f5c\u8005\u9009\u62e9\u4e86\u4e00\u4e2a\u73b0\u5b9e\u4e16\u754c\u4e2d\u7684\u65b9\u4f4d\u95ee\u9898\uff1a<\/p>\n

\"\u56fe\u7247\"<\/p>\n

GPT-4\u7b2c\u4e00\u6b21\u7ed9\u51fa\u7684\u7b54\u6848\u662f\u53f3\u8fb9\uff0c\u4f46\u4f5c\u8005\u6307\u51fa\u4e86\u5b83\u7684\u9519\u8bef\uff0c\u867d\u7136\u4ece\u5730\u56fe\u4e0a\u6765\u770b\uff0c\u4f4d\u4e8e\u9a6c\u8428\u8bf8\u585e\u5dde\u7684\u6ce2\u58eb\u987f\u7684\u786e\u5728\u5357\u8fbe\u79d1\u4ed6\u5dde\u7684\u53f3\u8fb9\uff0c\u4f46\u8fd9\u91cc\u8fd8\u6709\u4e00\u4e2a\u9644\u52a0\u6761\u4ef6\uff1a\u8eab\u4f53\u7684\u671d\u5411\u662f\u5f97\u514b\u8428\u65af\u5dde\u3002<\/p>\n

\"\u56fe\u7247\"<\/p>\n

\u8fd9\u610f\u5473\u7740\u6ce2\u58eb\u987f\u5728\u4f5c\u8005\u7684\u5de6\u8fb9\u3002<\/p>\n

\u4e4b\u540e\uff0cGPT-4\u5728\u56de\u7b54\u6ce2\u58eb\u987f\u4e0e\u5357\u8fbe\u79d1\u4ed6\u5dde\u9ad8\u4f4e\u4f4d\u7f6e\u65f6\uff0c\u51fa\u73b0\u4e86\u66f4\u4e25\u91cd\u7684\u95ee\u9898\uff1a\u5b83\u5728\u540c\u4e00\u4e2a\u56de\u7b54\u4e2d\u7ed9\u51fa\u4e86\u4e24\u79cd\u77db\u76fe\u7684\u63cf\u8ff0\u3002<\/p>\n

\"\u56fe\u7247\"<\/p>\n

13. \u65f6\u95f4\u63a8\u7406<\/strong><\/p>\n

\u4f5c\u8005\u5728\u8fd9\u91cc\u7ed9\u51fa\u4e86\u4e00\u4e2a\u6bd4\u8f83\u7b80\u5355\u7684\u65f6\u95f4\u63a8\u7406\u95ee\u9898\uff0c\u4f46GPT-4\u7684\u56de\u7b54\u4f9d\u65e7\u4e00\u584c\u7cca\u6d82\u3002<\/p>\n

\"\u56fe\u7247\"<\/p>\n

\n

Tom\u548cNancy\u4e0a\u73ed\u9700\u8981\u4e58\u5750\u4ea4\u901a\u5de5\u5177\u3002Nancy\u7684\u901a\u52e4\u65f6\u95f4\u5927\u7ea6\u4e3a30~40\u5206\u949f\uff0c\u800cTom\u7684\u901a\u52e4\u65f6\u95f4\u5927\u7ea6\u4e3a40~50\u5206\u949f\u3002\u4e0a\u4e2a\u5468\u4e94\uff0cNancy\u5728\u65e9\u4e0a8:10~8:20\u4e4b\u95f4\u79bb\u5bb6\uff0c\u800cTom\u5728\u65e9\u4e0a8:5~9:10\u4e4b\u95f4\u5230\u8fbe\u5de5\u4f5c\u5730\u70b9\u3002\u6b64\u5916\uff0cNancy\u5728Tom\u79bb\u5f00\u5bb6\u540e\u5230\u8fbe\u5de5\u4f5c\u5730\u70b9\uff0c\u4f46\u4e0d\u4f1a\u8d85\u8fc720\u5206\u949f\u3002\u4f60\u80fd\u5426\u63a8\u65ad\u51fa\u4e0a\u4e2a\u661f\u671f\u4e94\uff0cTom\u548cNancy\u4f55\u65f6\u5230\u8fbe\u5de5\u4f5c\u5730\u70b9\uff1f<\/p>\n<\/blockquote>\n

\u5728\u68b3\u7406\u5b8c\u95ee\u9898\u4e2d\u7684\u4fe1\u606f\u540e\uff0cGPT-4\u7ed9\u51fa\u4e86\u5b83\u7684\u63a8\u7406\u8fc7\u7a0b\uff1a<\/p>\n

\"\u56fe\u7247\"<\/p>\n

\u300c\u5982\u679cTom\u5728\u53ef\u80fd\u6700\u665a\u7684\u65f6\u95f4\uff08\u4e0a\u53488:20\uff09\u79bb\u5f00\u5bb6…\u300d \u8fd9\u53e5\u8bdd\u4e00\u5f00\u7bc7\u5c31\u9519\u4e86\u3002<\/p>\n

\u5b9e\u9645\u4e0a\uff0c\u9898\u76ee\u5e76\u6ca1\u6709\u7ed9\u51fa\u6709\u5173Tom\u6700\u665a\u79bb\u5f00\u5bb6\u7684\u65f6\u95f4\uff0c\u800cGPT-4\u5c06Nancy\u7684\u65f6\u95f4\uff08\u300cNancy\u5728\u4e0a\u53488:10-8:20\u4e4b\u95f4\u79bb\u5bb6\u300d\uff09\u8bef\u7528\u5230\u4e86Tom\u8eab\u4e0a\u3002<\/p>\n

\u540c\u65f6\uff0cGPT-4\u7ed9\u51fa\u7684\u6761\u4ef6\u8bed\u53e5\u662f\u6df7\u4e71\u7684\uff0c\u5047\u8bbe\u4e2d\u5305\u542b\u4e86\u4e0e\u7ed3\u8bba\uff08Nancy\u7684\u5230\u8fbe\u65f6\u95f4\uff09\u65e0\u5173\u7684\u4fe1\u606f\uff08Tom\uff09\uff1a\u300c\u5982\u679cTom\u5728\u6700\u665a\u65f6\u95f4\uff08\u4e0a\u53488:20\uff09\u79bb\u5f00\u5bb6\uff0cNancy\u5728\u5979\u6700\u665a\u65f6\u95f4\uff08\u4e0a\u53488:20\uff09\u79bb\u5f00\uff0c\u5979\u7684\u901a\u52e4\u65f6\u95f4\u6700\u591a\u662f40\u5206\u949f\uff0cNancy\u6700\u665a\u5728\u4e0a\u53489:00\u5230\u8fbe\u5de5\u4f5c\u5730\u70b9\u3002\u300d<\/p>\n

\u8fd9\u5e94\u8be5\u8868\u8ff0\u4e3a\uff1a\u300c\u5982\u679cNancy\u5728\u5979\u6700\u665a\u65f6\u95f4\uff08\u4e0a\u53488:20\uff09\u79bb\u5f00\uff0c\u5e76\u4e14\u5979\u7684\u901a\u52e4\u65f6\u95f4\u6700\u591a\u662f40\u5206\u949f\uff0c\u90a3\u4e48Nancy\u6700\u665a\u4f1a\u5728\u4e0a\u53489:00\u5230\u8fbe\u5de5\u4f5c\u5730\u70b9\u3002\u300d<\/p>\n

\u63a5\u7740\uff0cGPT-4\u9519\u8bef\u5730\u63a8\u65ad\u51fa\u4ee5\u4e0b\u5185\u5bb9\uff1a\u300c\u7531\u4e8eTom\u7684\u901a\u52e4\u65f6\u95f4\u6700\u5c11\u4e3a40\u5206\u949f\uff0c\u8fd9\u610f\u5473\u7740\u4ed6\u6700\u665a\u4f1a\u5728\u4e0a\u53489:00\u5230\u8fbe\u5de5\u4f5c\u5730\u70b9\u3002\u300d<\/p>\n

\u8fd9\u4e2a\u7ed3\u8bba\u663e\u800c\u6613\u89c1\u6839\u672c\u4e0d\u6210\u7acb\u3002\u4ece\u5df2\u77e5\u7684\u300cTom\u7684\u901a\u52e4\u65f6\u95f4\u6700\u5c11\u4e3a40\u5206\u949f\u300d\u8fd9\u4e2a\u4e8b\u5b9e\u4e2d\u65e0\u6cd5\u5f97\u51fa\u8fd9\u4e2a\u7ed3\u8bba\u3002<\/p>\n

\u63a5\u4e0b\u6765\u7684\u56de\u7b54\u4f9d\u65e7\u662f\u57fa\u4e8e\u9519\u8bef\u5730\u5047\u8bbeTom\u6700\u65e9\u79bb\u5f00\u65f6\u95f4\u662f\u4e0a\u53488:10\u7684\u6761\u4ef6\uff08\u518d\u6b21\uff0c\u8fd9\u4e2a\u51fa\u53d1\u65f6\u95f4\u662fNancy\u7684\uff0c\u4e0d\u662fTom\u7684\uff09\u3002<\/p>\n

\u7136\u540e\u5b83\u58f0\u79f0Nancy\u5230\u8fbe\u65f6\u95f4\u662f8:45\uff0c\u8fd9\u4e0e\u65e9\u4e0a8:10\u79bb\u5bb6\uff0c\u4e0d\u8d85\u8fc720\u5206\u949f\u6761\u4ef6\u4e0d\u7b26\u5408\u3002<\/p>\n

\u6700\u540e\uff0c\u5b83\u9519\u8bef\u5730\u5f97\u51fa\u7ed3\u8bbaTom\u548cNancy\u90fd\u57288:50\u548c9:00\u4e4b\u95f4\u5230\u8fbe\u3002<\/p>\n

\u5728\u63a8\u7406\u7684\u8fc7\u7a0b\u4e2d\uff0cGPT-4\u5c61\u6b21\u51fa\u73b0\u4e86\u5c06\u4fe1\u606f\u5f20\u51a0\u674e\u6234\u7684\u60c5\u51b5\uff0c\u6700\u540e\u7ed9\u51fa\u7684\u7b54\u6848\u4e5f\u662f\u57fa\u4e8e\u9519\u8bef\u6761\u4ef6\u5f97\u51fa\u7684\u9519\u8bef\u56de\u7b54\u3002<\/p>\n

14. \u8c0b\u6740\u8fd8\u662f\u81ea\u6740\uff1f<\/strong><\/p>\n

\u4f5c\u8005\u6784\u601d\u4e86\u4e00\u4e2a\u903b\u8f91\u8c1c\u9898\uff0c\u5217\u51fa\u4e869\u4e2a\u6761\u4ef6\u8981\u6c42GPT-4\u627e\u51fa\u771f\u6b63\u6740\u5bb3Agatha\u59e8\u5988\u7684\u51f6\u624b\u3002<\/p>\n

\"\u56fe\u7247\"<\/p>\n

\n

1. \u4f4f\u5728Dreadbury Mansion\u7684\u67d0\u4eba\u6740\u4e86Agatha\u59e8\u5988\u3002 <\/p>\n

2. Dreadbury Mansion\u4e2d\u552f\u4e00\u7684\u5c45\u4f4f\u8005\u662fAgatha\u59e8\u5988\u3001\u7ba1\u5bb6\u548cCharles\u3002 <\/p>\n

3. \u6740\u4eba\u72af\u603b\u662f\u8ba8\u538c\u4ed6\u7684\u53d7\u5bb3\u8005\uff0c\u5e76\u4e14\u4ed6\u7684\u8d22\u5bcc\u4e0d\u4f1a\u6bd4\u53d7\u5bb3\u8005\u591a\u3002<\/p>\n

4. Charles\u4e0d\u8ba8\u538cAgatha\u59e8\u5988\u8ba8\u538c\u7684\u4eba\u3002 <\/p>\n

5. Agatha\u59e8\u5988\u8ba8\u538c\u6240\u6709\u4eba\uff0c\u9664\u4e86\u7ba1\u5bb6\u3002 <\/p>\n

6. \u7ba1\u5bb6\u8ba8\u538c\u6240\u6709\u4e0d\u6bd4Agatha\u59e8\u5988\u5bcc\u6709\u7684\u4eba\u3002 <\/p>\n

7. \u7ba1\u5bb6\u8ba8\u538cAgatha\u59e8\u5988\u8ba8\u538c\u7684\u6240\u6709\u4eba\u3002<\/p>\n

8. \u6ca1\u6709\u4eba\u8ba8\u538c\u6240\u6709\u4eba\u3002 <\/p>\n

9. Agatha\u59e8\u5988\u4e0d\u662f\u7ba1\u5bb6\u3002<\/p>\n<\/blockquote>\n

\u6b63\u786e\u7684\u7b54\u6848\u662fAgatha\u59e8\u5988\u6740\u4e86\u81ea\u5df1\u3002<\/p>\n

\u9996\u5148\uff0c\u6839\u636e\u6761\u4ef65\uff0cAgatha\u59e8\u5988\u5fc5\u987b\u8ba8\u538c\u5979\u81ea\u5df1\uff0c\u56e0\u4e3a\u5979\u8ba8\u538c\u6240\u6709\u9664\u4e86\u7ba1\u5bb6\u4ee5\u5916\u7684\u4eba\u3002<\/p>\n

\u56e0\u6b64\uff0c\u6839\u636e\u6761\u4ef64\uff0c\u5f97\u51faCharles\u4e0d\u8ba8\u538c\u5979\uff0c\u6240\u4ee5\u4ed6\u4e0d\u53ef\u80fd\u6740\u4e86\u5979\u3002<\/p>\n

\u6839\u636e\u6761\u4ef65\u548c7\uff0c\u7ba1\u5bb6\u4e0d\u53ef\u80fd\u8ba8\u538c\u4ed6\u81ea\u5df1\uff0c\u56e0\u4e3a\u5982\u679c\u4ed6\u8ba8\u538c\u81ea\u5df1\u7684\u8bdd\uff0c\u6761\u4ef68\u5c31\u4e0d\u6210\u7acb\u4e86\uff0c\u4ed6\u4f1a\u8ba8\u538c\u6240\u6709\u4eba\u3002<\/p>\n

\u6839\u636e\u6761\u4ef66\uff0c\u5f97\u51fa\u7ba1\u5bb6\u6bd4Agatha\u59e8\u5988\u66f4\u5bcc\u6709\uff0c\u5426\u5219\u4ed6\u4f1a\u8ba8\u538c\u81ea\u5df1\uff0c\u8fd9\u4e0e\u524d\u9762\u6211\u4eec\u5f97\u51fa\u7684\u4ed6\u4e0d\u8ba8\u538c\u81ea\u5df1\u76f8\u77db\u76fe\u3002<\/p>\n

\u6839\u636e\u6761\u4ef63\uff0c\u7ba1\u5bb6\u4e5f\u4e0d\u4f1a\u662f\u51f6\u624b\uff08\u7b2c3\u4e2a\u6761\u4ef6\uff09\u3002<\/p>\n

\"\u56fe\u7247\"<\/p>\n

\u5728\u63a8\u7406\u4e2d\uff0cGPT-4\u6b63\u786e\u5730\u6392\u9664\u4e86Charles\uff0c\u4f46\u65e0\u6cd5\u6392\u9664\u7ba1\u5bb6\uff0c\u5e76\u5f97\u51fa\u4e86\u9519\u8bef\u7684\u7ed3\u8bba\uff1a\u7ba1\u5bb6\u662f\u51f6\u624b\u3002<\/p>\n

GPT-4\u505a\u51fa\u7684\u53e6\u4e00\u4e2a\u5173\u952e\u9519\u8bef\u662f\uff1a\u7531\u4e8eAgatha\u59e8\u5988\u8ba8\u538c\u6240\u6709\u9664\u7ba1\u5bb6\u4ee5\u5916\u7684\u4eba\uff08\u6761\u4ef65\uff09\uff0c\u8fd9\u610f\u5473\u7740\u5979\u81f3\u5c11\u4e0d\u8ba8\u538c\u5979\u81ea\u5df1\u3002<\/p>\n

\u8fd9\u662f\u4e00\u4e2a\u5947\u602a\u7684\u9519\u8bef\uff0c\u4ece\u7b2c5\u4e2a\u6761\u4ef6\u5c31\u53ef\u4ee5\u5f97\u51faAgatha\u59e8\u5988\u8ba8\u538c\u5979\u81ea\u5df1\u3002<\/p>\n

\u540c\u65f6\uff0cGPT-4\u53c8\u4e00\u6b21\u5c55\u793a\u4e86\u53cd\u590d\u51fa\u73b0\u7684\u4e0d\u4e00\u81f4\u6027\u95ee\u9898\u2014\u2014\u51e0\u4e4e\u5728\u6bcf\u4e00\u6761\u56de\u590d\u4e2d\uff0cGPT-4\u90fd\u58f0\u79f0\u63a8\u5bfc\u51fa\u67d0\u4e2a\u547d\u9898\u53ca\u5176\u5426\u5b9a\u5f62\u5f0f\u3002<\/p>\n

15. \u6c83\u68ee\u9009\u62e9\u4efb\u52a1\uff08Wason selection task\uff09<\/strong><\/p>\n

\u6c83\u68ee\u9009\u62e9\u4efb\u52a1\u662f\u5fc3\u7406\u63a8\u7406\u9886\u57df\u4e2d\u7684\u57fa\u672c\u5185\u5bb9\u3002<\/p>\n

\u5728\u4e00\u6708\u4efd\u7684\u8bba\u6587\u4e2d\uff0cGPT-3.5\u5c31\u672a\u80fd\u901a\u8fc7\u8fd9\u4e2a\u6d4b\u8bd5\uff0c\u672c\u6b21\u7814\u7a76\u4e2d\uff0cGPT-4\u7684\u8868\u73b0\u4f9d\u65e7\u4e0d\u7406\u60f3\u3002<\/p>\n

\"\u56fe\u7247\"<\/p>\n

\n

\u684c\u4e0a\u653e\u77407\u5f20\u724c\uff0c\u6bcf\u5f20\u724c\u4e00\u9762\u5199\u7740\u6570\u5b57\uff0c\u53e6\u4e00\u9762\u662f\u5355\u8272\u8272\u5757\u3002\u8fd9\u4e9b\u724c\u7684\u6b63\u9762\u663e\u793a\u7684\u662f50\u300116\u3001\u7ea2\u8272\u3001\u9ec4\u8272\u300123\u3001\u7eff\u8272\u300130\u3002<\/p>\n

\u8981\u5224\u65ad\u300c\u5982\u679c\u4e00\u5f20\u724c\u6b63\u9762\u663e\u793a4\u7684\u500d\u6570\uff0c\u5219\u80cc\u9762\u989c\u8272\u4e3a\u9ec4\u8272\u300d\u8fd9\u4e2a\u547d\u9898\u7684\u771f\u5047\uff0c\u4f60\u9700\u8981\u7ffb\u8f6c\u54ea\u4e9b\u724c?<\/p>\n<\/blockquote>\n

\"\u56fe\u7247\"<\/p>\n

\u8fd9\u4e9b\u56de\u7b54\u663e\u793a\uff0cGPT-4\u4e0d\u7406\u89e3\u6761\u4ef6\u8bed\u53e5\u7684\u8bed\u4e49\u3002\u5f53GPT-4\u8bf4\u5361\u7247\u300c50\u300d\u548c\u300c30\u300d\u5fc5\u987b\u7ffb\u5f00\u65f6\uff0c\u5b83\u4f3c\u4e4e\u5c06\u6761\u4ef6\u8bef\u8ba4\u4e3a\u662f\u5145\u5206\u5fc5\u8981\u6761\u4ef6\u3002<\/p>\n

\u800c\u65e0\u8bbaGPT-4\u7684\u56de\u7b54\u662f\u5bf9\u8fd8\u662f\u9519\uff0c\u5176\u5185\u90e8\u7684\u8bf4\u6cd5\u90fd\u662f\u4e0d\u4e00\u81f4\u7684\u3002<\/p>\n

16. \u71b5<\/strong><\/p>\n

\u4fe1\u606f\u8bba\u7684\u4e00\u4e2a\u57fa\u672c\u7ed3\u8bba\u662f\uff1a\u968f\u673a\u5411\u91cfZ\u7684\u71b5\u4e0a\u754c\u4e0d\u8d85\u8fc7\u7ec4\u6210Z\u7684\u968f\u673a\u53d8\u91cf\u7684\u71b5\u4e4b\u548c\u3002<\/p>\n

\u56e0\u6b64\uff0c\u4e0b\u9762\u95ee\u9898\u7684\u7b54\u6848\u5e94\u8be5\u662f\u300c\u5728\u4efb\u4f55\u60c5\u51b5\u4e0b\u90fd\u4e0d\u4f1a\u300d\u3002<\/p>\n

\"\u56fe\u7247\"<\/p>\n

17. \u7b80\u5355\u7f16\u8bd1\u5668\u7684\u6b63\u786e\u6027<\/strong><\/p>\n

\u6700\u540e\u7ed9GPT-4\u7684\u63a8\u7406\u95ee\u9898\u662f\u6700\u5177\u6311\u6218\u6027\u7684\uff1a\u8bc1\u660e\u4e00\u4e2a\u7b80\u5355\u8868\u8fbe\u5f0f\u7f16\u8bd1\u5668\u7684\u6b63\u786e\u6027\u3002<\/p>\n

\"\u56fe\u7247\"<\/p>\n

<\/p>\n

\u4f46\u5728\u8fd9\u6b21\u6d4b\u8bd5\u4e2d\uff0cGPT-4\u901a\u8fc7\u5728\u8868\u8fbe\u5f0f\u7684\u62bd\u8c61\u8bed\u6cd5\u7ed3\u6784\u4e0a\u8bbe\u7f6e\u7ed3\u6784\u5f52\u7eb3\uff0c\u6b63\u786e\u5730\u8fdb\u884c\u4e86\u8bc1\u660e\u3002<\/p>\n

\u8fd9\u53ef\u80fd\u662f\u56e0\u4e3a\u5b83\u4e4b\u524d\u770b\u8fc7\u7c7b\u4f3c\u7684\u8bc1\u660e\uff0c\u4f5c\u8005\u7ed9\u51fa\u7684\u4f8b\u5b50\u662f\u7f16\u7a0b\u8bfe\u7a0b\u548c\u6559\u6750\u4e2d\u5e38\u89c1\u7684\u7ec3\u4e60\u7c7b\u578b\u3002<\/p>\n

\u7136\u800c\uff0cGPT-4\u8fd8\u662f\u4f1a\u51fa\u73b0\u4e00\u4e9b\u7ec6\u8282\u4e0a\u9519\u8bef\u3002<\/p>\n

\"\u56fe\u7247\"<\/p>\n

\u7ed3\u8bba\uff1a\u63a8\u7406\u80fd\u529b\u81f3\u5173\u91cd\u8981\uff0c\u4f46GPT-4\u4e0d\u4f1a<\/strong><\/p>\n

\u9274\u4e8eGPT-4\u662f\u76ee\u524d\u80fd\u529b\u6700\u5f3a\u7684LLM\uff0c\u56e0\u6b64\u4f5c\u8005\u57fa\u4e8e\u4ee5\u4e0a\u5206\u6790\u7ed9\u51fa\u4e86\u4e09\u4e2a\u4e3b\u8981\u7ed3\u8bba\uff1a<\/p>\n

1. \u5728\u8f6f\u4ef6\u5f00\u53d1\uff08\u6216\u4e00\u822c\u7684\u79d1\u5b66\u548c\u5de5\u7a0b\u9886\u57df\uff09\u4e2d\u4f7f\u7528\u751f\u6210\u5f0fAI\uff0c\u9664\u4e86\u5bf9\u4e8e\u4e00\u4e9b\u7e41\u7410\u7684\u4efb\u52a1\u5916\uff08\u4f5c\u4e3a\u4e00\u79cd\u5bf9\u77e5\u8bc6\u5bc6\u96c6\u578b\u7f16\u7801\u95ee\u9898\u7684\u52a0\u901f\u81ea\u52a8\u8865\u5168\uff09\uff0c\u5145\u6ee1\u4e86\u98ce\u9669\u3002\u5728\u8fd9\u4e9b\u9886\u57df\uff0c\u89c4\u8303\u6027\u548c\u6b63\u786e\u6027\u81f3\u5173\u91cd\u8981\uff0c\u800c\u5f53\u524d\u7684LLM\u65e0\u6cd5\u8fbe\u5230\u8fd9\u4e9b\u6807\u51c6\u3002<\/p>\n

2. \u968f\u7740LLM\u63a8\u7406\u80fd\u529b\u7684\u4e0d\u65ad\u63d0\u9ad8\uff0c\u4e25\u683c\u7684\u8bc1\u660e\u68c0\u67e5\u4f1a\u53d8\u5f97\u8d8a\u6765\u8d8a\u91cd\u8981\u3002\u8fd9\u79cd\u65b9\u6cd5\u53ef\u4ee5\u901a\u8fc7\u8981\u6c42LLM\u5c06\u5176\u63a8\u7406\u5f62\u5f0f\u5316\uff0c\u6216\u8005\u901a\u8fc7\u8bad\u7ec3\u5176\u4ed6LLM\uff0c\u6765\u68c0\u67e5\u7528\u81ea\u7136\u8bed\u8a00\u8868\u8fbe\u7684\u63a8\u7406\u3002<\/p>\n

3. \u5c31\u76ee\u524d\u800c\u8a00\uff0cAI\u5f81\u670d\u4eba\u7c7b\u6216\u4eba\u7c7b\u5229\u7528AI\u8fbe\u5230\u90aa\u6076\u76ee\u7684\u8fd9\u79cd\u53cd\u4e4c\u6258\u90a6\u60c5\u666f\uff0c\u90fd\u6781\u4e3a\u7275\u5f3a\uff0c\u751a\u81f3\u5230\u4e86\u8352\u8c2c\u7684\u5730\u6b65\u3002\u5f53\u6700\u5148\u8fdb\u7684AI\u7cfb\u7edf\u8fde\u5de6\u53f3\u90fd\u5206\u4e0d\u6e05\u65f6\uff08\u4e0a\u8ff0\u7b2c12\u4e2a\u95ee\u9898\uff09\uff0c\u547c\u5401\u5236\u5b9a\u653f\u7b56\u6765\u4fdd\u62a4\u4eba\u7c7b\u514d\u53d7\u5b83\u7684\u4f24\u5bb3\uff0c\u5f80\u597d\u91cc\u8bf4\u662f\u4e3a\u65f6\u8fc7\u65e9\uff0c\u5f80\u5927\u4e86\u8bf4\u5c31\u662f\u5bf9\u8d44\u6e90\u7684\u6d6a\u8d39\u3002<\/p>\n

\u4e0d\u53ef\u907f\u514d\u5730\uff0c\u4e00\u4e9b\u4eba\u53ef\u80fd\u4f1a\u8bf4\u8fd9\u4e9b\u7ed3\u679c\u662f\u300c\u6311\u9009\u6570\u636e\u300d\u3002\u4f46\u8fd9\u662f\u56e0\u4e3a\u4ed6\u4eec\u5bf9\u4ec0\u4e48\u662f\u6311\u9009\u6570\u636e\u5b58\u5728\u7740\u8bef\u89e3\u3002\u6839\u636e\u76f8\u5173\u547d\u9898\u7684\u903b\u8f91\u7ed3\u6784\u548c\u6574\u4f53\u80cc\u666f\uff0c\u6311\u9009\u6570\u636e\u6709\u65f6\u751a\u81f3\u662f\u5fc5\u8981\u7684\u3002<\/p>\n

\u901a\u8fc7\u5bf9\u8ba1\u7b97\u673a\u7a0b\u5e8f\u8fdb\u884c\u8c03\u8bd5\u6765\u53d1\u73b0\u548c\u7406\u89e3\u5176\u5f31\u70b9\uff0c\u8bd5\u56fe\u8bc1\u4f2a\u79d1\u5b66\u7406\u8bba\uff0c\u8bd5\u9a7e\u65b0\u8f66\uff0c\u8bd5\u56fe\u627e\u5230\u4e00\u4e2a\u5047\u5b9a\u7684\u5b9a\u7406\u7684\u53cd\u6a21\u578b\u7b49\u7b49\uff0c\u4ece\u6839\u672c\u4e0a\u6765\u8bf4\u90fd\u662f\u300c\u6311\u523a\u300d\u3002<\/p>\n

\u4e3e\u4e2a\u4f8b\u5b50\uff0c\u6bd4\u5982\u4f60\u53d1\u73b0\u81ea\u5df1\u65b0\u4e70\u7684\u6c7d\u8f66\u6709\u4e00\u4e2a\u8f6e\u80ce\u6f0f\u6c14\uff0c\u8fd9\u65f6\u7ecf\u9500\u5546\u5c31\u53ef\u4ee5\u6297\u8bae\u79f0\u4f60\u662f\u5728\u300c\u6311\u9009\u6570\u636e\u300d\u3002\u6bd5\u7adf\uff0c\u5c31\u6574\u8f86\u8f66\u6765\u8bf4\uff0c\u8f6e\u80ce\u7684\u5b8c\u597d\u7387\u9ad8\u8fbe75%\u3002<\/p>\n

\u540c\u6837\uff0c\u79d1\u5b66\u3001\u533b\u5b66\u548c\u5de5\u7a0b\u9886\u57df\u7684\u5e94\u7528\uff0c\u5c24\u5176\u662f\u8f6f\u4ef6\u5de5\u7a0b\uff0c\u90fd\u6709\u4e25\u683c\u7684\u6807\u51c6\u3002<\/p>\n

\u5c31\u50cf\u6211\u4eec\u4e0d\u60f3\u8981\u4e00\u5ea7\u572890%\u7684\u60c5\u51b5\u4e0b\u80fd\u7acb\u67f1\u7684\u6865\u6881\u4e00\u6837\uff0c\u6211\u4eec\u9700\u8981\u5bf9\u6240\u6709\u8f93\u5165\u90fd\u6709\u6548\u7684\u6392\u5e8f\u7b97\u6cd5\uff0c\u800c\u4e0d\u4ec5\u4ec5\u662f\u5927\u90e8\u5206\uff1b\u6211\u4eec\u9700\u8981\u8d2d\u7269\u8f66\u6bcf\u6b21\u90fd\u80fd\u6536\u53d6\u6b63\u786e\u7684\u8d39\u7528\uff0c\u800c\u4e0d\u4ec5\u4ec5\u662f\u5927\u591a\u6570\u65f6\u95f4\uff0c\u7b49\u7b49\u3002<\/p>\n

\u800c\u8fd9\u4e9b\u8ba1\u7b97\u548c\u63a8\u7406\u5bc6\u96c6\u578b\u7684\u5e94\u7528\uff0c\u4e0e\u63a8\u8350\u5f15\u64ce\u4e0d\u540c\uff0c\u5b83\u4eec\u5fc5\u987b\u975e\u5e38\u53ef\u9760\u3002<\/p>\n

\u4f5c\u8005\u4ecb\u7ecd<\/strong><\/p>\n

Konstantine Arkoudas<\/strong><\/p>\n

\u76f4\u5230\u53bb\u5e74\uff0cKonstantine Arkoudas\u8fd8\u662fRPI\u8ba4\u77e5\u79d1\u5b66\u7cfb\u7684\u7814\u7a76\u5458\uff0c\u4e5f\u662f\u9ebb\u7701\u7406\u5de5\u5b66\u9662CSAIL\u7684\u7814\u7a76\u5458\u3002<\/p>\n

\u76ee\u524d\uff0c\u4ed6\u662fTelcordia\u7814\u7a76\u5b9e\u9a8c\u5ba4\u7684\u9ad8\u7ea7\u7814\u7a76\u79d1\u5b66\u5bb6\uff0c\u4e3b\u8981\u7814\u7a76AI\uff0c\u4ee5\u53ca\u5728\u7535\u4fe1\u548c\u7f51\u7edc\u884c\u4e1a\u5e94\u7528\u6b63\u5f0f\u65b9\u6cd5\u89e3\u51b3\u73b0\u5b9e\u4e16\u754c\u7684\u95ee\u9898\u3002<\/p>\n

\u4ed6\u66fe\u57282000\u5e74\u83b7\u5f97\u4e86MIT\u7684\u8ba1\u7b97\u673a\u79d1\u5b66\u535a\u58eb\u5b66\u4f4d\u3002\u5728\u6b64\u4e4b\u524d\uff0c\u8fd8\u83b7\u5f97\u4e86\u8ba1\u7b97\u673a\u79d1\u5b66\u7855\u58eb\u5b66\u4f4d\uff0c\u54f2\u5b66\u7855\u58eb\u5b66\u4f4d\uff0c\u4ee5\u53ca\u8ba1\u7b97\u673a\u79d1\u5b66\u5b66\u58eb\u5b66\u4f4d\uff0c\u8f85\u4fee\u54f2\u5b66\u3002<\/p>\n

\u5927\u5b66\u6570\u7406\u5316\uff0cGPT-4\u5f97\u520635.8%<\/h2>\n

UCLA\u7684\u7814\u7a76\u4e2d\uff0c\u4e3b\u8981\u8bc4\u4f30\u4e86GPT-4\uff0c\u4ee5\u53caGPT-3.5\u5728\u6570\u5b66\u3001\u5316\u5b66\u3001\u7269\u7406\u65b9\u9762\u7684\u63a8\u7406\u80fd\u529b\u3002<\/p>\n

\u5f53\u524d\uff0c\u4e3a\u4e86\u589e\u5f3aLLM\u89e3\u51b3\u6570\u5b66\u7b49\u4efb\u52a1\u7684\u80fd\u529b\uff0c\u6709\u4eba\u63d0\u51fa\u4e86\u601d\u7ef4\u8fdeCoT\u7b56\u7565\uff0c\u6307\u5bfc\u5927\u6a21\u578b\u9010\u6b65\u751f\u6210\u7b54\u6848\uff0c\u4ece\u800c\u66f4\u6df1\u5165\u601d\u8003\u95ee\u9898\u3002<\/p>\n

\u7136\u800c\uff0c\u5373\u4f7f\u8fd9\u6837\u7684\u65b9\u6cd5\u6709\u5176\u7279\u5b9a\u7684\u4f18\u52bf\uff0c\u4e5f\u96be\u4ee5\u5b8c\u5168\u89e3\u51b3\u590d\u6742\u7684\u79d1\u5b66\u95ee\u9898\u3002<\/p>\n

\u5982\u4e0b\uff0c\u662f\u5927\u5b66\u7269\u7406\u5316\u5b66\u7684\u4e00\u4e2a\u793a\u4f8b\u95ee\u9898\uff0c\u4ee5\u53ca\u5728\u4e24\u79cd\u63d0\u793a\u7b56\u7565\u4e0b\u751f\u6210\u7684\u89e3\u51b3\u65b9\u6848\u3002<\/p>\n

\u6709CoT\u52a0\u6301\u7684GPT-4\u51fa\u73b0\u660e\u663e\u7684\u8ba1\u7b97\u9519\u8bef\uff0c\u800c\u63d0\u793a\u7528Python\u4f5c\u4e3a\u5916\u90e8\u5de5\u5177\u7684GPT-4\uff0c\u4e5f\u4f1a\u8bef\u89e3\u6570\u5b66\u65b9\u7a0b\u3002<\/p>\n

\"\u56fe\u7247\"<\/p>\n

\u9519\u8bef\u6807\u8bb0\u4e3a\u7ea2\u8272\uff0c\u66f4\u6b63\u5185\u5bb9\u4e3a\u7d2b\u8272<\/p>\n

\u5bf9\u6b64\uff0c\u7814\u7a76\u4e2d\u5f15\u5165\u4e86\u4e00\u4e2a\u5927\u5b66\u6c34\u5e73\u7684\u79d1\u5b66\u95ee\u9898\u57fa\u51c6SCIBENCH\u3002<\/p>\n

\u5176\u4e2d\uff0c\u300c\u5f00\u653e\u6570\u636e\u96c6\u300d\u5305\u62ec\u4ece\u5927\u5b66\u8bfe\u7a0b\u5e7f\u6cdb\u4f7f\u7528\u7684\u6559\u79d1\u4e66\u4e2d\u6536\u96c6\u76845\u4e2a\u95ee\u9898\uff0c\u6db5\u76d6\u4e86\u57fa\u7840\u7269\u7406\u3001\u70ed\u529b\u5b66\u3001\u7ecf\u5178\u529b\u5b66\u3001\u91cf\u5b50\u5316\u5b66\u3001\u7269\u7406\u5316\u5b66\u3001\u5fae\u79ef\u5206\u3001\u7edf\u8ba1\u5b66\u548c\u5fae\u5206\u65b9\u7a0b\u3002<\/p>\n

\"\u56fe\u7247\"<\/p>\n

\u5f00\u653e\u6559\u79d1\u4e66\u95ee\u9898\u6458\u8981\uff08\u5305\u62ec\u95ee\u9898\u6570\u91cf\u7684\u6bd4\u4f8b\uff0c\u4ee5\u53ca\u6709\u8be6\u7ec6\u89e3\u51b3\u65b9\u6848\u7684\u6bd4\u4f8b\uff09<\/p>\n

\u53e6\u4e00\u4e2a\u662f\u300c\u5c01\u95ed\u6570\u636e\u96c6\u300d\uff0c\u4e3a\u4e86\u6a21\u62df\u771f\u5b9e\u4e16\u754c\u7684\u8bc4\u4f30\uff0c\u5176\u4e2d\u5305\u542b\u4e86\u8ba1\u7b97\u673a\u79d1\u5b66\u548c\u6570\u5b66\u4e09\u95e8\u5927\u5b66\u8bfe\u7a0b\u76847\u5957\u671f\u4e2d\u548c\u671f\u672b\u8003\u8bd5\u9898\u3002<\/p>\n

\"\u56fe\u7247\"<\/p>\n

\u5c01\u95ed\u8003\u8bd5\u6570\u636e\u96c6\uff08\u5305\u542b\u6bcf\u573a\u8003\u8bd5\u4e2d\u7684\u95ee\u9898\u5b9e\u4f8b\u6570\uff0c\u4ee5\u53ca\u8003\u8bd5\u4e2d\u5305\u542b\u8be6\u7ec6\u89e3\u7b54\u7684\u95ee\u9898\u6bd4\u4f8b\u3002\u53e6\u5916\uff0c\u8fd8\u6709\u4e0d\u540c\u5f62\u5f0f\u95ee\u9898\u7684\u6bd4\u4f8b\uff0c\u5305\u62ec\u81ea\u7531\u56de\u7b54\u3001\u591a\u9879\u9009\u62e9\u548c\u771f\u5047\u7b54\u6848\u3002\u4f5c\u4e3a\u53c2\u8003\uff0c\u62ec\u53f7\u4e2d\u7684\u6570\u5b57\u8868\u793a\u95ee\u9898\u7684\u8bc4\u5206\u70b9\u3002\uff09<\/p>\n

\u4e0e\u73b0\u6709\u57fa\u51c6\u4e0d\u540c\uff0cSCIBENCH\u4e2d\u7684\u6240\u6709\u95ee\u9898\u90fd\u662f\uff0c\u5f00\u653e\u5f0f\u3001\u81ea\u7531\u56de\u7b54\u7684\u95ee\u9898\u3002<\/p>\n

\u6570\u636e\u96c6\u4e2d\u6709\u4e86\uff0c\u7814\u7a76\u91cd\u70b9\u8bc4\u4f30\u4e86\u4e24\u4e2a\u5177\u6709\u4ee3\u8868\u6027\u7684LLM\uff0cGPT-3.5\u548cGPT-4\uff0c\u5e76\u91c7\u7528\u4e86\u4e0d\u540c\u7684\u63d0\u793a\u7b56\u7565\uff0c\u5305\u62ecCoT\u3001\u96f6\u6837\u672c\u5b66\u4e60\u3001\u5c11\u6837\u672c\u5b66\u4e60\u3002<\/p>\n

\u53e6\u5916\uff0c\u7814\u7a76\u4eba\u5458\u8fd8\u63d0\u793a\u6a21\u578b\u4f7f\u7528\u5916\u90e8\u5de5\u5177\uff0c\u6bd4\u5982Python\u548cWolfram\u8bed\u8a00\u3002<\/p>\n

\u5b9e\u9a8c\u7ed3\u679c\u8868\u660e\uff0c\u5728\u6ca1\u6709\u4efb\u4f55\u590d\u6742\u63d0\u793a\u3001\u6216\u4f7f\u7528\u5916\u90e8\u5de5\u5177\u7684\u60c5\u51b5\u4e0b\uff0cGPT-3.5\u548cGPT-4\u5728\u5f00\u653e\u6570\u636e\u96c6\u4e2d\u5e73\u5747\u51c6\u786e\u7387\u5206\u522b\u4e3a10.62%\u548c16.81%\u3002<\/p>\n

\u90a3\u4e48\uff0c\u5728\u52a0\u5165CoT\u548c\u5916\u90e8\u5de5\u5177\u540e\uff0c\u5728\u540c\u4e00\u6570\u636e\u96c6\u4e0a\u6700\u9ad8\u51c6\u786e\u7387\u4e5f\u4ec5\u4ec5\u662f35.8%\u3002\u4e0d\u8fc7\uff0c\u76f8\u8f83\u4e4b\u524d\uff0c\u5f88\u5927\u7a0b\u5ea6\u63d0\u9ad8\u4e86\u51c6\u786e\u7387\u3002<\/p>\n

\"\u56fe\u7247\"<\/p>\n

\u5f00\u653e\u6570\u636e\u96c6\u4e2d\u51c6\u786e\u7387\u7684\u7ed3\u679c<\/p>\n

\u5728\u4f7f\u7528CoT\u63d0\u793a+\u5916\u90e8\u5de5\u5177\u6700\u5f3a\u914d\u7f6e\u4e0b\uff0cGPT-4\u5728\u5f00\u653e\u5f0f\u6570\u636e\u96c6\u4e0a\u53d6\u5f97\u4e8635.80%\u7684\u5e73\u5747\u5206\uff0c\u5728\u5c01\u95ed\u6570\u636e\u96c6\u4e0a\u53d6\u5f97\u4e8651.57%\u7684\u5e73\u5747\u5206\u3002<\/p>\n

\u8fd9\u4e9b\u7ed3\u679c\u8868\u660e\uff0c\u5728\u672a\u6765\u7684LLM\u4e2d\uff0cGPT-4\u6709\u76f8\u5f53\u5927\u7684\u6539\u8fdb\u6f5c\u529b\u3002<\/p>\n

\"\u56fe\u7247\"<\/p>\n

\u8003\u8bd5\u6570\u636e\u96c6\u4e0a\u96f6\u6837\u672c\u5b66\u4e60\u4e0b\u603b\u5206\u7684\u5b9e\u9a8c\u7ed3\u679c<\/p>\n

\u4e3a\u4e86\u5168\u9762\u4e86\u89e3LLM\u5728\u79d1\u5b66\u95ee\u9898\u89e3\u51b3\u4e2d\u7684\u5c40\u9650\u6027\uff0c\u7814\u7a76\u4eba\u5458\u63d0\u51fa\u4e86\u4e00\u79cd\u5168\u65b0\u7684\u300c\u81ea\u6211\u5b8c\u5584\u300d\u7684\u65b9\u6cd5\uff0c\u4ee5\u53d1\u73b0LLM\u6240\u505a\u89e3\u7b54\u4e2d\u7684\u4e0d\u8db3\u4e4b\u5904\u3002<\/p>\n

\u4fbf\u662f\u5982\u4e0b\u7684\u300c\u8bc4\u4f30\u534f\u8bae\u300d\u3002<\/p>\n

\"\u56fe\u7247\"<\/p>\n

\u9996\u5148\uff0c\u5c06\u6b63\u786e\u7684\u89e3\u51b3\u65b9\u6848\u4e0eLLM\u751f\u6210\u7684\u89e3\u51b3\u65b9\u6848\u8fdb\u884c\u6bd4\u8f83\uff0c\u5e76\u5728\u4eba\u5de5\u6807\u6ce8\u5458\u7684\u534f\u52a9\u4e0b\uff0c\u603b\u7ed3\u51fa\u6210\u529f\u89e3\u51b3\u79d1\u5b66\u95ee\u9898\u6240\u9700\u768410\u9879\u57fa\u672c\u6280\u80fd\u3002<\/p>\n

\u5177\u4f53\u5305\u62ec\uff1a\u903b\u8f91\u5206\u89e3\u548c\u5206\u6790\u80fd\u529b\uff1b\u8bc6\u522b\u5047\u8bbe\uff1b\u7a7a\u95f4\u611f\u77e5\uff1b\u56e0\u679c\u63a8\u7406\uff1b\u95ee\u9898\u6f14\u7ece\uff1b\u62bd\u8c61\u63a8\u7406\uff1b\u79d1\u5b66\u7d20\u517b\uff1b\u4ee3\u7801\u8f6c\u6362\uff1b\u903b\u8f91\u63a8\u7406\uff1b\u8ba1\u7b97\u80fd\u529b\u3002<\/p>\n

\u968f\u540e\uff0c\u56e2\u961f\u91c7\u7528\u4e86\u4e00\u79cd\u7531LLM\u9a71\u52a8\u7684\u81ea\u6211\u8bc4\u4ef7\u65b9\u6cd5\uff0c\u5bf9\u6bcf\u4e2a\u5b9e\u9a8c\u914d\u7f6e\u4e0b\u57fa\u51c6LLM\u6240\u505a\u7684\u89e3\u51b3\u65b9\u6848\u4e2d\uff0c\u7f3a\u4e4f\u7684\u6280\u80fd\u8fdb\u884c\u81ea\u52a8\u5206\u7c7b\u3002<\/p>\n

\"\u56fe\u7247\"<\/p>\n

6\u79cd\u8bbe\u7f6e\u4e0bGPT-3.5\u5728\u6587\u672c\u6570\u636e\u96c6\u4e0a\u7684\u9519\u8bef\u6982\u51b5\uff0c\u63ed\u793a\u4e86\u517610\u79cd\u57fa\u672c\u89e3\u51b3\u95ee\u9898\u80fd\u529b\u7684\u7f3a\u9677\u5206\u5e03<\/p>\n

\u6700\u540e\uff0c\u901a\u8fc7\u5206\u6790\u53d1\u73b0\uff1a<\/p>\n

(1) \u867d\u7136CoT\u663e\u8457\u63d0\u9ad8\u4e86\u8ba1\u7b97\u80fd\u529b\uff0c\u4f46\u5728\u5176\u4ed6\u65b9\u9762\u7684\u6548\u679c\u8f83\u5dee\uff1b<\/p>\n

(2) \u4f7f\u7528\u5916\u90e8\u5de5\u5177\u7684\u63d0\u793a\u53ef\u80fd\u4f1a\u635f\u5bb3\u5176\u4ed6\u57fa\u672c\u6280\u80fd\uff1b<\/p>\n

(3) \u5c11\u6837\u672c\u5b66\u4e60\u5e76\u4e0d\u80fd\u666e\u904d\u63d0\u9ad8\u79d1\u5b66\u95ee\u9898\u89e3\u51b3\u80fd\u529b\u3002<\/p>\n

\u603b\u4e4b\uff0c\u7814\u7a76\u7ed3\u679c\u8868\u660e\uff0c\u5f53\u524d\u5927\u578b\u8bed\u8a00\u6a21\u578b\u5728\u89e3\u51b3\u95ee\u9898\u80fd\u529b\u65b9\u9762\u4f9d\u65e7\u5f88\u5f31\uff0c\u5e76\u4e14\u5728\u5404\u79cd\u5de5\u5177\u5e2e\u52a9\u4e0b\uff0c\u4f9d\u65e7\u5b58\u5728\u5c40\u9650\u6027\u3002<\/p>\n

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

https:\/\/www.preprints.org\/manuscript\/202308.0148\/v2<\/p>\n

https:\/\/arxiv.org\/pdf\/2307.10635.pdf<\/p>\n","protected":false},"excerpt":{"rendered":"

\u7f16\u8005\u6309\uff1a\u672c\u6587\u6765\u81ea\u5fae\u4fe1\u516c\u4f17\u53f7 \u65b0\u667a\u5143\uff08ID\uff1aAI_era\uff09\uff0c\u7f16\u8f91\uff1a\u7f16\u8f91...<\/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-2970","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\/2970","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=2970"}],"version-history":[{"count":0,"href":"http:\/\/www.idea2003.com\/wp-json\/wp\/v2\/posts\/2970\/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=2970"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/www.idea2003.com\/wp-json\/wp\/v2\/categories?post=2970"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/www.idea2003.com\/wp-json\/wp\/v2\/tags?post=2970"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}