{"id":5596,"date":"2025-01-25T12:20:05","date_gmt":"2025-01-25T04:20:05","guid":{"rendered":"https:\/\/nullthought.net\/?p=5596"},"modified":"2025-01-25T13:38:46","modified_gmt":"2025-01-25T05:38:46","slug":"mind-evolution%ef%bc%9a%e7%bb%93%e5%90%88%e5%9f%ba%e5%9b%a0%e6%90%9c%e7%b4%a2%e7%ae%97%e6%b3%95%e4%b8%8ellm%e7%9a%84%e7%94%9f%e6%88%90%e8%83%bd%e5%8a%9b%ef%bc%8c%e5%88%a9%e7%94%a8%e8%bf%9b%e5%8c%96","status":"publish","type":"post","link":"https:\/\/nullthought.net\/?p=5596","title":{"rendered":"Mind Evolution\uff1a\u7ed3\u5408\u57fa\u56e0\u641c\u7d22\u7b97\u6cd5\u4e0eLLM\u7684\u751f\u6210\u80fd\u529b\uff0c\u5229\u7528\u8fdb\u5316\u7b56\u7565\u5728\u81ea\u7136\u8bed\u8a00\u7a7a\u95f4\u4e2d\u9ad8\u6548\u641c\u7d22\u89e3\u51b3\u65b9\u6848"},"content":{"rendered":"\n<p>\u8bba\u6587<strong><a href=\"https:\/\/arxiv.org\/abs\/2501.09891\" target=\"_blank\" rel=\"noreferrer noopener\">Evolving Deeper LLM Thinking<\/a><\/strong>\u63d0\u51fa\u4e86\u4e00\u79cd\u521b\u65b0\u7684\u8fdb\u5316\u641c\u7d22\u7b56\u7565\u2014\u2014Mind Evolution\uff0c\u65e8\u5728\u901a\u8fc7\u66f4\u52a0\u9ad8\u6548\u5730\u5229\u7528\u63a8\u7406\u65f6\u95f4\u8ba1\u7b97\u8d44\u6e90\uff0c\u63d0\u5347\u5927\u578b\u8bed\u8a00\u6a21\u578b\uff08LLM\uff09\u5728\u89e3\u51b3\u590d\u6742\u95ee\u9898\u65f6\u7684\u63a8\u7406\u80fd\u529b\u3002\u4f20\u7edf\u7684LLM\u89e3\u51b3\u95ee\u9898\u65b9\u6cd5\u5f80\u5f80\u4f9d\u8d56\u4e8e\u5355\u4e00\u7684\u63a8\u7406\u6b65\u9aa4\u6216\u56fa\u5b9a\u7684\u63a8\u7406\u8def\u5f84\uff0c\u800cMind Evolution\u901a\u8fc7\u5f15\u5165\u57fa\u56e0\u641c\u7d22\u7684\u601d\u60f3\uff0c\u7ed3\u5408\u81ea\u7531\u6d41\u52a8\u7684\u968f\u673a\u63a2\u7d22\u4e0e\u5927\u89c4\u6a21\u7684\u8fed\u4ee3\u4f18\u5316\uff0c\u5bfb\u6c42\u66f4\u52a0\u7075\u6d3b\u548c\u9ad8\u6548\u7684\u89e3\u51b3\u65b9\u6848\u3002<\/p>\n\n\n\n<p>\u8bba\u6587\u4f5c\u8005\u4e3aKuang-Huei Lee, Ian Fischer, Yueh-Hua Wu, Dave Marwood, Shumeet Baluja, Dale Schuurmans, Xinyun Chen\uff0c\u6765\u81eaGoogle DeepMind, UC San Diego, University of Alberta\u3002<\/p>\n\n\n\n<figure class=\"wp-block-image aligncenter size-full\"><img loading=\"lazy\" decoding=\"async\" width=\"902\" height=\"466\" src=\"https:\/\/nullthought.net\/wp-content\/uploads\/2025\/01\/image-26.png\" alt=\"\" class=\"wp-image-5597\" srcset=\"https:\/\/nullthought.net\/wp-content\/uploads\/2025\/01\/image-26.png 902w, https:\/\/nullthought.net\/wp-content\/uploads\/2025\/01\/image-26-300x155.png 300w, https:\/\/nullthought.net\/wp-content\/uploads\/2025\/01\/image-26-768x397.png 768w\" sizes=\"auto, (max-width: 902px) 100vw, 902px\" \/><figcaption class=\"wp-element-caption\"><strong><a href=\"https:\/\/arxiv.org\/abs\/2501.09891\" target=\"_blank\" rel=\"noreferrer noopener\">Evolving Deeper LLM Thinking<\/a><\/strong><\/figcaption><\/figure>\n\n\n\n<h5 class=\"wp-block-heading\">\u4e00\u3001\u7814\u7a76\u80cc\u666f\u4e0e\u95ee\u9898\u6982\u8ff0<\/h5>\n\n\n\n<p>\u5728\u8fc7\u53bb\u7684\u7814\u7a76\u4e2d\uff0c\u6709\u591a\u79cd\u7b56\u7565\u7528\u4e8e\u5229\u7528\u63a8\u7406\u8ba1\u7b97\u8d44\u6e90\uff0c\u5982<a href=\"https:\/\/nullthought.net\/?p=5067\" target=\"_blank\" rel=\"noreferrer noopener\">\u94fe\u5f0f\u601d\u7ef4\uff08Chain-of-Thought\uff09<\/a>\u3001\u81ea\u4e00\u81f4\u6027\uff08Self-consistency\uff09\u3001\u53cd\u9988\u9a71\u52a8\u7684\u9010\u6b65\u4fee\u6b63\uff08Sequential Revision\uff09\u3001\u4ee5\u53ca\u4f9d\u8d56\u8f85\u52a9\u9a8c\u8bc1\u5668\u6216\u8bc4\u4f30\u5668\u7684\u641c\u7d22\u7b56\u7565\uff08\u5982Best-of-N\u548c\u6811\u641c\u7d22\uff09\u3002\u8fd9\u4e9b\u65b9\u6cd5\u901a\u5e38\u4f9d\u8d56\u4e8e\u660e\u786e\u7684\u63a8\u7406\u6b65\u9aa4\u6216\u542f\u53d1\u5f0f\u7684\u53cd\u9988\u8fc7\u7a0b\u6765\u9010\u6b65\u6539\u5584\u89e3\u7b54\uff0c\u800cMind Evolution\u5219\u901a\u8fc7\u4e00\u5957\u5168\u65b0\u7684\u8fdb\u5316\u7b97\u6cd5\uff0c\u5c06\u95ee\u9898\u89e3\u7684\u751f\u6210\u4e0e\u53cd\u9988\u7ed3\u5408\uff0c\u907f\u514d\u4e86\u9700\u8981\u660e\u786e\u5f62\u5f0f\u5316\u95ee\u9898\u7684\u9700\u6c42\uff0c\u6781\u5927\u5730\u6269\u5c55\u4e86\u5176\u5e94\u7528\u573a\u666f\u3002<\/p>\n\n\n\n<p>\u5177\u4f53\u6765\u8bf4\uff0cMind Evolution\u901a\u8fc7\u4ee5\u4e0b\u4e24\u4e2a\u91cd\u8981\u7279\u70b9\u533a\u5206\u4e8e\u5176\u4ed6\u65b9\u6cd5\uff1a<\/p>\n\n\n\n<ol class=\"wp-block-list\">\n<li><strong>\u81ea\u7531\u63a2\u7d22\u4e0e\u7cbe\u7ec6\u8c03\u6574\u7684\u7ed3\u5408<\/strong>\uff1a\u8be5\u65b9\u6cd5\u4e0d\u4ec5\u8fdb\u884c\u5e7f\u6cdb\u7684\u63a2\u7d22\uff08\u7c7b\u4f3c\u4e8eBest-of-N\uff09\uff0c\u540c\u65f6\u8fd8\u5728\u591a\u6b21\u8fed\u4ee3\u4e2d\u5bf9\u6700\u6709\u6f5c\u529b\u7684\u89e3\u8fdb\u884c\u6df1\u5ea6\u8c03\u6574\uff0c\u4ece\u800c\u63d0\u9ad8\u4e86\u627e\u5230\u6700\u4f73\u89e3\u7684\u53ef\u80fd\u6027\u3002<\/li>\n\n\n\n<li><strong>\u65e0\u9700\u9010\u6b65\u8bc4\u4f30\u8fc7\u7a0b<\/strong>\uff1a\u4e0d\u540c\u4e8e\u6811\u641c\u7d22\u7b49\u9010\u6b65\u63a8\u7406\u65b9\u6cd5\uff0cMind Evolution\u91c7\u7528\u5168\u7403\u6027\u7684\u89e3\u7684\u4f18\u5316\u7b56\u7565\uff0c\u53ea\u9700\u8981\u4e00\u4e2a\u5168\u5c40\u7684\u89e3\u8bc4\u4f30\u5668\uff0c\u800c\u4e0d\u5fc5\u9010\u4e2a\u8bc4\u4f30\u63a8\u7406\u6b65\u9aa4\u3002<\/li>\n<\/ol>\n\n\n\n<p>\u6b64\u5916\uff0cMind Evolution\u91c7\u7528\u7684\u8fdb\u5316\u641c\u7d22\u65b9\u6cd5\u80fd\u591f\u5e76\u884c\u5904\u7406\u5927\u91cf\u89e3\u5019\u9009\uff0c\u9002\u5408\u5728\u591a\u6838\u8ba1\u7b97\u73af\u5883\u4e0b\u5b9e\u73b0\u9ad8\u6548\u8ba1\u7b97\u3002<\/p>\n\n\n\n<h5 class=\"wp-block-heading\">\u4e8c\u3001\u65b9\u6cd5\u4e0e\u6280\u672f\u6846\u67b6<\/h5>\n\n\n\n<p>Mind Evolution\u7ed3\u5408\u4e86\u57fa\u56e0\u641c\u7d22\u7b97\u6cd5\u4e0eLLM\u7684\u751f\u6210\u80fd\u529b\uff0c\u6838\u5fc3\u601d\u60f3\u662f\u5229\u7528\u8fdb\u5316\u7b56\u7565\u5728\u81ea\u7136\u8bed\u8a00\u7a7a\u95f4\u4e2d\u9ad8\u6548\u641c\u7d22\u89e3\u51b3\u65b9\u6848\u3002\u6574\u4e2a\u8fc7\u7a0b\u5305\u62ec\u5019\u9009\u751f\u6210\u3001\u9002\u5e94\u5ea6\u8bc4\u4f30\u3001\u4ea4\u53c9\u4e0e\u53d8\u5f02\u64cd\u4f5c\uff0c\u4ee5\u53ca\u591a\u4ee3\u8fed\u4ee3\u4f18\u5316\uff0c\u6700\u7ec8\u901a\u8fc7\u5168\u7403\u8bc4\u4f30\u9009\u62e9\u6700\u4f73\u89e3\u51b3\u65b9\u6848\u3002<\/p>\n\n\n\n<h6 class=\"wp-block-heading\">2.1 \u57fa\u56e0\u641c\u7d22\u7b97\u6cd5\u6982\u8ff0<\/h6>\n\n\n\n<p><a href=\"https:\/\/en.wikipedia.org\/wiki\/Genetic_algorithm\" target=\"_blank\" rel=\"noreferrer noopener\">\u57fa\u56e0\u7b97\u6cd5\uff08Genetic Algorithm\uff0cGA\uff09<\/a>\u662f\u4e00\u79cd\u6a21\u4eff\u81ea\u7136\u9009\u62e9\u8fc7\u7a0b\u7684\u641c\u7d22\u4f18\u5316\u7b97\u6cd5\uff0c\u80fd\u591f\u901a\u8fc7\u6a21\u62df\u751f\u7269\u8fdb\u5316\u8fc7\u7a0b\u4e2d\u7684\u201c\u9009\u62e9\u3001\u4ea4\u53c9\u4e0e\u53d8\u5f02\u201d\u6765\u5bfb\u627e\u95ee\u9898\u7684\u6700\u4f18\u89e3\u3002\u5728Mind Evolution\u4e2d\uff0c\u57fa\u56e0\u641c\u7d22\u7684\u6bcf\u4e00\u4ee3\u90fd\u4f1a\u5bf9\u5019\u9009\u89e3\u8fdb\u884c\u8bc4\u5206\uff0c\u9009\u62e9\u8868\u73b0\u8f83\u597d\u7684\u89e3\u8fdb\u884c\u4ea4\u53c9\u4e0e\u53d8\u5f02\uff0c\u751f\u6210\u65b0\u7684\u5019\u9009\u89e3\uff0c\u8fd9\u4e00\u8fc7\u7a0b\u4e0d\u65ad\u91cd\u590d\uff0c\u76f4\u5230\u627e\u5230\u6ee1\u8db3\u8981\u6c42\u7684\u89e3\u6216\u8fbe\u5230\u8ba1\u7b97\u9884\u7b97\u3002<\/p>\n\n\n\n<p>Mind Evolution\u7684\u5b9e\u73b0\u6d41\u7a0b\u5982\u4e0b\uff1a<\/p>\n\n\n\n<ol class=\"wp-block-list\">\n<li><strong>\u5019\u9009\u751f\u6210<\/strong>\uff1a\u901a\u8fc7LLM\u751f\u6210\u521d\u59cb\u89e3\u3002\u5728\u95ee\u9898\u63cf\u8ff0\u548c\u76f8\u5173\u63d0\u793a\u7684\u5e2e\u52a9\u4e0b\uff0cLLM\u53ef\u4ee5\u751f\u6210\u591a\u4e2a\u5019\u9009\u89e3\u3002<\/li>\n\n\n\n<li><strong>\u9002\u5e94\u5ea6\u8bc4\u4f30<\/strong>\uff1a\u4f7f\u7528\u9884\u8bbe\u7684\u8bc4\u4f30\u51fd\u6570\u5bf9\u5019\u9009\u89e3\u8fdb\u884c\u6253\u5206\u3002\u8bc4\u4f30\u51fd\u6570\u6839\u636e\u5019\u9009\u89e3\u662f\u5426\u6ee1\u8db3\u7ed9\u5b9a\u7684\u7ea6\u675f\u6761\u4ef6\uff08\u5982\u9884\u7b97\u3001\u65f6\u95f4\u3001\u4f4d\u7f6e\u7b49\uff09\u4ee5\u53ca\u5176\u6574\u4f53\u8d28\u91cf\uff08\u5982\u7b26\u5408\u5ea6\u3001\u521b\u65b0\u6027\u7b49\uff09\u6765\u4e3a\u89e3\u8d4b\u503c\u3002<\/li>\n\n\n\n<li><strong>\u4ea4\u53c9\u4e0e\u53d8\u5f02<\/strong>\uff1a\u901a\u8fc7\u4ea4\u53c9\u64cd\u4f5c\u5c06\u591a\u4e2a\u9ad8\u9002\u5e94\u5ea6\u89e3\u8fdb\u884c\u7ec4\u5408\uff0c\u53d8\u5f02\u64cd\u4f5c\u5219\u968f\u673a\u6539\u53d8\u89e3\u7684\u90e8\u5206\u5185\u5bb9\uff0c\u4ee5\u589e\u52a0\u89e3\u7684\u591a\u6837\u6027\u3002<\/li>\n\n\n\n<li><strong>\u8fed\u4ee3\u4f18\u5316<\/strong>\uff1a\u8be5\u8fc7\u7a0b\u4f1a\u5728\u591a\u4ee3\u4e4b\u95f4\u4e0d\u65ad\u8fed\u4ee3\uff0c\u901a\u8fc7\u7cbe\u7ec6\u7684\u4ea4\u53c9\u4e0e\u53d8\u5f02\u64cd\u4f5c\u4ea7\u751f\u66f4\u52a0\u4f18\u79c0\u7684\u89e3\uff0c\u9010\u6b65\u4f18\u5316\u95ee\u9898\u7684\u89e3\u51b3\u65b9\u6848\u3002<\/li>\n<\/ol>\n\n\n\n<h6 class=\"wp-block-heading\">2.2 \u63d0\u5347\u6279\u5224\u6027\u601d\u7ef4\u7684RCC\u8fc7\u7a0b<\/h6>\n\n\n\n<p>Mind Evolution\u5f15\u5165\u4e86\u201c\u6279\u5224\u6027\u5bf9\u8bdd\u201d\uff08Refinement through Critical Conversation\uff0cRCC\uff09\u8fc7\u7a0b\uff0c\u8fd9\u662f\u63d0\u5347\u6a21\u578b\u6279\u5224\u6027\u601d\u7ef4\u7684\u5173\u952e\u6240\u5728\u3002\u5728\u6b64\u8fc7\u7a0b\u4e2d\uff0cLLM\u901a\u8fc7\u201c\u6279\u8bc4\u8005\u201d\u548c\u201c\u4f5c\u8005\u201d\u4e24\u4e2a\u89d2\u8272\u8fdb\u884c\u4ea4\u66ff\u5de5\u4f5c\uff1a<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>\u6279\u8bc4\u8005\u89d2\u8272<\/strong>\uff1a\u5206\u6790\u5019\u9009\u89e3\uff0c\u5e76\u6839\u636e\u8bc4\u4f30\u7ed3\u679c\u7ed9\u51fa\u6539\u8fdb\u5efa\u8bae\uff0c\u6307\u51fa\u89e3\u7684\u7f3a\u9677\u6216\u4e0d\u6ee1\u8db3\u7684\u7ea6\u675f\u3002<\/li>\n\n\n\n<li><strong>\u4f5c\u8005\u89d2\u8272<\/strong>\uff1a\u6839\u636e\u6279\u8bc4\u8005\u7684\u53cd\u9988\u8fdb\u884c\u6539\u8fdb\uff0c\u63d0\u51fa\u65b0\u7684\u89e3\uff0c\u5e76\u7ee7\u7eed\u63d0\u4ea4\u7ed9\u6279\u8bc4\u8005\u8fdb\u884c\u5206\u6790\u3002<\/li>\n<\/ul>\n\n\n\n<p>\u8fd9\u79cd\u6279\u5224\u6027\u601d\u7ef4\u7684\u5206\u79bb\u5e2e\u52a9\u6a21\u578b\u8fdb\u884c\u6df1\u5ea6\u81ea\u6211\u53cd\u7701\uff0c\u4f7f\u5f97\u89e3\u7684\u8d28\u91cf\u9010\u6b65\u63d0\u9ad8\u3002RCC\u8fc7\u7a0b\u7c7b\u4f3c\u4e8e\u201c\u53cd\u601d\u201d\u7b97\u6cd5\uff08Reflexion\uff09\uff0c\u901a\u8fc7\u591a\u8f6e\u5bf9\u8bdd\u7684\u65b9\u5f0f\u6539\u8fdb\u89e3\uff0c\u4f7f\u5176\u66f4\u52a0\u7b26\u5408\u95ee\u9898\u8981\u6c42\u3002<\/p>\n\n\n\n<h6 class=\"wp-block-heading\">2.3 \u5c9b\u5c7f\u6a21\u578b\u4e0e\u79cd\u7fa4\u591a\u6837\u6027<\/h6>\n\n\n\n<p>Mind Evolution\u91c7\u7528\u5c9b\u5c7f\u6a21\u578b\uff08Island Model\uff09\u6765\u7ef4\u62a4\u79cd\u7fa4\u7684\u591a\u6837\u6027\uff0c\u907f\u514d\u8fdb\u5316\u8fc7\u7a0b\u4e2d\u7684\u8fc7\u5ea6\u6536\u655b\u3002\u5177\u4f53\u800c\u8a00\uff0c\u5c9b\u5c7f\u6a21\u578b\u5c06\u89e3\u7684\u79cd\u7fa4\u5206\u4e3a\u591a\u4e2a\u5b50\u7fa4\u4f53\uff0c\u8fd9\u4e9b\u5b50\u7fa4\u4f53\u4e4b\u95f4\u8fdb\u884c\u5468\u671f\u6027\u7684\u8fc1\u79fb\u548c\u91cd\u7f6e\uff0c\u4ece\u800c\u4fdd\u6301\u79cd\u7fa4\u7684\u591a\u6837\u6027\uff0c\u907f\u514d\u9677\u5165\u5c40\u90e8\u6700\u4f18\u89e3\u3002\u6bcf\u4e2a\u5c9b\u5c7f\u4e0a\u7684\u89e3\u96c6\u5728\u4e00\u5b9a\u6570\u91cf\u7684\u4ee3\u6570\u540e\u4f1a\u8fdb\u884c\u8fc1\u79fb\uff0c\u5373\u4f18\u8d28\u89e3\u4ece\u4e00\u4e2a\u5c9b\u5c7f\u8fc1\u79fb\u5230\u76f8\u90bb\u5c9b\u5c7f\uff0c\u901a\u8fc7\u8fd9\u79cd\u65b9\u5f0f\u52a0\u901f\u6574\u4e2a\u8fdb\u5316\u8fc7\u7a0b\u3002<\/p>\n\n\n\n<h6 class=\"wp-block-heading\">2.4 \u5168\u5c40\u4f18\u5316\u4e0e\u5e76\u884c\u5316<\/h6>\n\n\n\n<p>Mind Evolution\u7684\u4f18\u5316\u8fc7\u7a0b\u662f\u5168\u5c40\u6027\u7684\uff0c\u610f\u5473\u7740\u5b83\u901a\u8fc7\u4e00\u6b21\u6027\u8bc4\u4f30\u5b8c\u6574\u89e3\u800c\u975e\u5355\u6b65\u63a8\u7406\uff0c\u8fd9\u4f7f\u5f97\u8be5\u65b9\u6cd5\u80fd\u591f\u66f4\u9ad8\u6548\u5730\u4f7f\u7528\u63a8\u7406\u8ba1\u7b97\u8d44\u6e90\u3002\u6b64\u5916\uff0c\u8be5\u65b9\u6cd5\u9002\u5408\u5e76\u884c\u5316\uff0c\u53ef\u4ee5\u5728\u591a\u6838\u8ba1\u7b97\u73af\u5883\u4e0b\u5e76\u884c\u751f\u6210\u548c\u8bc4\u4f30\u591a\u4e2a\u89e3\uff0c\u663e\u8457\u63d0\u5347\u8ba1\u7b97\u6548\u7387\u3002<\/p>\n\n\n\n<h5 class=\"wp-block-heading\">\u4e09\u3001\u5b9e\u9a8c\u8bbe\u8ba1\u4e0e\u7ed3\u679c\u5206\u6790<\/h5>\n\n\n\n<p>\u8bba\u6587\u5728\u591a\u4e2a\u81ea\u7136\u8bed\u8a00\u89c4\u5212\u4efb\u52a1\u4e0a\u9a8c\u8bc1\u4e86Mind Evolution\u7684\u6709\u6548\u6027\uff0c\u5c24\u5176\u5728TravelPlanner\u3001Trip Planning\u548cMeeting Planning\u8fd9\u4e09\u4e2a\u4efb\u52a1\u4e2d\u5c55\u73b0\u4e86\u660e\u663e\u7684\u4f18\u52bf\u3002<\/p>\n\n\n\n<h6 class=\"wp-block-heading\">3.1 \u65c5\u884c\u8ba1\u5212\uff08TravelPlanner\uff09<\/h6>\n\n\n\n<p><strong>\u4efb\u52a1\u63cf\u8ff0<\/strong>\uff1aTravelPlanner\u662f\u4e00\u4e2a\u81ea\u7136\u8bed\u8a00\u89c4\u5212\u57fa\u51c6\u4efb\u52a1\uff0c\u8981\u6c42\u6a21\u578b\u751f\u6210\u6ee1\u8db3\u7528\u6237\u9700\u6c42\u7684\u65c5\u884c\u8ba1\u5212\u3002\u95ee\u9898\u901a\u8fc7\u81ea\u7136\u8bed\u8a00\u63cf\u8ff0\u4e00\u7cfb\u5217\u7684\u7ea6\u675f\u6761\u4ef6\uff0c\u5982\u9884\u7b97\u3001\u9910\u996e\u504f\u597d\u7b49\uff0c\u4efb\u52a1\u7684\u76ee\u6807\u662f\u751f\u6210\u7b26\u5408\u8fd9\u4e9b\u7ea6\u675f\u6761\u4ef6\u7684\u65c5\u884c\u8ba1\u5212\u3002<\/p>\n\n\n\n<p><strong>\u5b9e\u9a8c\u7ed3\u679c<\/strong>\uff1a\u5728TravelPlanner\u4efb\u52a1\u4e2d\uff0cMind Evolution\u8868\u73b0\u51fa\u4e86\u538b\u5012\u6027\u7684\u4f18\u52bf\u3002\u4e0e\u4f20\u7edf\u7684Best-of-N\u65b9\u6cd5\uff08\u6210\u529f\u738755.6%\uff09\u548cSequential Revision\u65b9\u6cd5\uff08\u6210\u529f\u738782.8%\uff09\u76f8\u6bd4\uff0cMind Evolution\u8fbe\u5230\u4e8695.6%\u7684\u6210\u529f\u7387\uff0c\u51e0\u4e4e\u89e3\u51b3\u4e86\u6240\u6709\u7684\u9a8c\u8bc1\u96c6\u95ee\u9898\u3002\u5f53\u4f7f\u7528\u66f4\u5f3a\u5927\u7684Gemini 1.5 Pro\u6a21\u578b\u65f6\uff0cMind Evolution\u6210\u529f\u7387\u8fbe\u5230\u4e86100%\u3002<\/p>\n\n\n\n<h6 class=\"wp-block-heading\">3.2 \u81ea\u7136\u8ba1\u5212\u2014\u2014\u65c5\u884c\u89c4\u5212\uff08Trip Planning\uff09<\/h6>\n\n\n\n<p><strong>\u4efb\u52a1\u63cf\u8ff0<\/strong>\uff1aTrip Planning\u4efb\u52a1\u8981\u6c42\u751f\u6210\u4e00\u4e2a\u7b26\u5408\u65f6\u95f4\u548c\u57ce\u5e02\u7ea6\u675f\u7684\u65c5\u884c\u884c\u7a0b\uff0c\u786e\u4fdd\u7528\u6237\u80fd\u591f\u5728\u6307\u5b9a\u7684\u65f6\u95f4\u548c\u5730\u70b9\u5b8c\u6210\u65c5\u884c\uff0c\u5e76\u4e14\u6240\u6709\u7684\u7ea6\u675f\u6761\u4ef6\u5f97\u5230\u6ee1\u8db3\u3002<\/p>\n\n\n\n<p><strong>\u5b9e\u9a8c\u7ed3\u679c<\/strong>\uff1a\u5728\u6b64\u4efb\u52a1\u4e2d\uff0cMind Evolution\u540c\u6837\u8d85\u8d8a\u4e86\u57fa\u7ebf\u65b9\u6cd5\uff0c\u9a8c\u8bc1\u96c6\u6210\u529f\u7387\u4e3a96.2%\uff0c\u6d4b\u8bd5\u96c6\u6210\u529f\u7387\u4e3a94.1%\u3002\u7279\u522b\u662f\uff0cMind Evolution\u5728\u5904\u7406\u591a\u57ce\u5e02\u89c4\u5212\u65f6\uff0c\u6bd4Best-of-N\u65b9\u6cd5\uff08\u6210\u529f\u738777.2%\uff09\u548cSequential Revision\u65b9\u6cd5\uff08\u6210\u529f\u738774.4%\uff09\u8868\u73b0\u5f97\u66f4\u52a0\u7a33\u5b9a\u548c\u9ad8\u6548\u3002<\/p>\n\n\n\n<h6 class=\"wp-block-heading\">3.3 \u81ea\u7136\u8ba1\u5212\u2014\u2014\u4f1a\u8bae\u5b89\u6392\uff08Meeting Planning\uff09<\/h6>\n\n\n\n<p><strong>\u4efb\u52a1\u63cf\u8ff0<\/strong>\uff1a\u5728\u4f1a\u8bae\u5b89\u6392\u4efb\u52a1\u4e2d\uff0c\u6a21\u578b\u9700\u8981\u5b89\u6392\u591a\u4e2a\u4f1a\u8bae\uff0c\u6700\u5927\u5316\u4e0e\u5c3d\u53ef\u80fd\u591a\u7684\u4e0e\u4f1a\u8005\u7684\u4f1a\u8bae\u6570\u91cf\uff0c\u4e14\u6ee1\u8db3\u65f6\u95f4\u3001\u5730\u70b9\u7b49\u7ea6\u675f\u3002<\/p>\n\n\n\n<p><strong>\u5b9e\u9a8c\u7ed3\u679c<\/strong>\uff1aMind Evolution\u5728\u4f1a\u8bae\u5b89\u6392\u4efb\u52a1\u4e2d\u7684\u8868\u73b0\u4e5f\u975e\u5e38\u7a81\u51fa\u3002\u9a8c\u8bc1\u96c6\u6210\u529f\u7387\u4e3a85.0%\uff0c\u6d4b\u8bd5\u96c6\u6210\u529f\u7387\u4e3a83.8%\u3002\u4f7f\u7528Gemini 1.5 Pro\u6a21\u578b\u7684\u4e24\u9636\u6bb5\u65b9\u6cd5\u540e\uff0c\u6210\u529f\u7387\u63d0\u5347\u81f398.4%\u3002\u76f8\u6bd4\u4e4b\u4e0b\uff0cBest-of-N\u65b9\u6cd5\u4ec5\u8fbe\u5230\u4e8669.4%\u7684\u6210\u529f\u7387\uff0cSequential Revision\u65b9\u6cd5\u7684\u6210\u529f\u7387\u4e3a62.0%\u3002<\/p>\n\n\n\n<h6 class=\"wp-block-heading\">3.4 \u65b0\u4efb\u52a1\u2014\u2014\u9690\u79d8\u8bd7\u6b4c\uff08StegPoet\uff09<\/h6>\n\n\n\n<p><strong>\u4efb\u52a1\u63cf\u8ff0<\/strong>\uff1aStegPoet\u662f\u4e00\u4e2a\u65b0\u5f15\u5165\u7684\u4efb\u52a1\uff0c\u8981\u6c42\u5c06\u9690\u85cf\u7684\u4fe1\u606f\u7f16\u7801\u5230\u4e00\u7bc7\u521b\u610f\u6587\u672c\uff08\u5982\u8bd7\u6b4c\uff09\u4e2d\uff0c\u6587\u672c\u9700\u8981\u9075\u5faa\u7279\u5b9a\u7684\u4e3b\u9898\u548c\u7ed3\u6784\uff0c\u5e76\u786e\u4fdd\u4fe1\u606f\u7684\u9690\u79d8\u6027\u3002<\/p>\n\n\n\n<p><strong>\u5b9e\u9a8c\u7ed3\u679c<\/strong>\uff1a\u5728StegPoet\u4efb\u52a1\u4e2d\uff0cMind Evolution\u7684\u9a8c\u8bc1\u96c6\u6210\u529f\u7387\u4e3a46.5%\uff0c\u4f7f\u7528Gemini 1.5 Pro\u6a21\u578b\u65f6\u63d0\u5347\u81f387.1%\u3002\u4e0e\u57fa\u7ebf\u65b9\u6cd5\u76f8\u6bd4\uff0cMind Evolution\u80fd\u591f\u5728\u6b64\u590d\u6742\u4efb\u52a1\u4e2d\u5c55\u793a\u51fa\u5176\u641c\u7d22\u548c\u4f18\u5316\u7684\u6f5c\u529b\uff0c\u5c3d\u7ba1\u8fd9\u4e2a\u4efb\u52a1\u5f88\u96be\u5f62\u5f0f\u5316\u3002<\/p>\n\n\n\n<h5 class=\"wp-block-heading\">\u56db\u3001\u7ec4\u4ef6\u5206\u6790\u4e0e\u6d88\u878d\u5b9e\u9a8c<\/h5>\n\n\n\n<p>\u901a\u8fc7\u4e00\u7cfb\u5217\u7684\u6d88\u878d\u5b9e\u9a8c\uff0c\u4f5c\u8005\u63a2\u8ba8\u4e86Mind Evolution\u5404\u4e2a\u7ec4\u4ef6\u7684\u4f5c\u7528\uff0c\u63ed\u793a\u4e86\u6279\u5224\u5bf9\u8bdd\uff08RCC\uff09\u8fc7\u7a0b\u5728\u63d0\u5347\u89e3\u8d28\u91cf\u65b9\u9762\u7684\u91cd\u8981\u6027\u3002\u5177\u4f53\u6765\u8bf4\uff1a<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>\u6279\u5224\u6027\u5206\u6790\u6b65\u9aa4<\/strong>\uff1a\u7981\u7528\u6279\u5224\u8005\u6b65\u9aa4\u4f1a\u663e\u8457\u964d\u4f4eMind Evolution\u7684\u6027\u80fd\uff08\u4ece95.6%\u964d\u81f346.1%\uff09\u3002\u8fd9\u8bc1\u660e\u4e86\u6279\u5224\u6027\u5206\u6790\u5bf9\u4e8e\u6a21\u578b\u6027\u80fd\u7684\u81f3\u5173\u91cd\u8981\u6027\u3002<\/li>\n\n\n\n<li><strong>\u5c9b\u5c7f\u6a21\u578b<\/strong>\uff1a\u4f7f\u7528\u5c9b\u5c7f\u6a21\u578b\u663e\u8457\u63d0\u9ad8\u4e86\u89e3\u7684\u591a\u6837\u6027\u548c\u7b97\u6cd5\u7684\u6536\u655b\u901f\u5ea6\u3002\u7981\u7528\u5c9b\u5c7f\u6a21\u578b\u65f6\uff0c\u6027\u80fd\u4e0b\u964d\u4e8610%\u3002<\/li>\n\n\n\n<li><strong>\u6587\u672c\u53cd\u9988<\/strong>\uff1a\u6587\u672c\u53cd\u9988\u7684\u52a0\u5165\u5927\u5e45\u63d0\u5347\u4e86\u89e3\u7684\u4f18\u5316\u8d28\u91cf\uff0c\u53cd\u9988\u673a\u5236\u80fd\u591f\u5e2e\u52a9\u6a21\u578b\u66f4\u597d\u5730\u7406\u89e3\u7ea6\u675f\u6761\u4ef6\u5e76\u8fdb\u884c\u76f8\u5e94\u8c03\u6574\u3002<\/li>\n<\/ul>\n\n\n\n<h5 class=\"wp-block-heading\">\u4e94\u3001\u7814\u7a76\u610f\u4e49\u4e0e\u5c40\u9650\u6027<\/h5>\n\n\n\n<h6 class=\"wp-block-heading\">5.1 \u7814\u7a76\u610f\u4e49<\/h6>\n\n\n\n<p>Mind Evolution\u7684\u5f15\u5165\u4e3a\u81ea\u7136\u8bed\u8a00\u89c4\u5212\u95ee\u9898\u7684\u89e3\u51b3\u63d0\u4f9b\u4e86\u4e00\u4e2a\u65b0\u7684\u65b9\u5411\uff0c\u7279\u522b\u662f\u5728\u65e0\u9700\u660e\u786e\u5f62\u5f0f\u5316\u95ee\u9898\u7684\u60c5\u51b5\u4e0b\uff0c\u5b83\u80fd\u591f\u5728\u591a\u4e2a\u590d\u6742\u4efb\u52a1\u4e2d\u8868\u73b0\u51fa\u5353\u8d8a\u7684\u6027\u80fd\u3002\u8fd9\u4e00\u65b9\u6cd5\u4e0d\u4ec5\u5bf9\u81ea\u7136\u8bed\u8a00\u5904\u7406\uff08NLP\uff09\u9886\u57df\u7684\u53d1\u5c55\u6709\u91cd\u8981\u610f\u4e49\uff0c\u4e5f\u4e3a\u4f18\u5316\u63a8\u7406\u8ba1\u7b97\u8d44\u6e90\u63d0\u4f9b\u4e86\u65b0\u7684\u89c6\u89d2\u3002<\/p>\n\n\n\n<h6 class=\"wp-block-heading\">5.2 \u5c40\u9650\u6027<\/h6>\n\n\n\n<p>\u5c3d\u7ba1Mind Evolution\u5728\u81ea\u7136\u8bed\u8a00\u89c4\u5212\u4efb\u52a1\u4e2d\u8868\u73b0\u4f18\u5f02\uff0c\u4f46\u76ee\u524d\u8be5\u65b9\u6cd5\u4ecd\u7136\u5c40\u9650\u4e8e\u80fd\u591f\u901a\u8fc7\u7a0b\u5e8f\u5316\u8bc4\u4f30\u7684\u4efb\u52a1\u3002\u672a\u6765\u7684\u7814\u7a76\u5c06\u81f4\u529b\u4e8e\u5f00\u53d1\u66f4\u901a\u7528\u7684LLM\u8bc4\u4f30\u5668\uff0c\u4ee5\u6269\u5c55\u5176\u5e94\u7528\u8303\u56f4\uff0c\u5e76\u63a2\u7d22\u5176\u4ed6\u7c7b\u578b\u4efb\u52a1\u7684\u4f18\u5316\u53ef\u80fd\u6027\u3002<\/p>\n","protected":false},"excerpt":{"rendered":"<p>\u8bba\u6587Evolving Deeper LLM Thinking\u63d0\u51fa\u4e86\u4e00\u79cd\u521b\u65b0\u7684\u8fdb\u5316\u641c\u7d22\u7b56\u7565\u2014\u2014Mind Evol [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"site-sidebar-layout":"default","site-content-layout":"","ast-site-content-layout":"default","site-content-style":"default","site-sidebar-style":"default","ast-global-header-display":"","ast-banner-title-visibility":"","ast-main-header-display":"","ast-hfb-above-header-display":"","ast-hfb-below-header-display":"","ast-hfb-mobile-header-display":"","site-post-title":"","ast-breadcrumbs-content":"","ast-featured-img":"","footer-sml-layout":"","ast-disable-related-posts":"","theme-transparent-header-meta":"","adv-header-id-meta":"","stick-header-meta":"","header-above-stick-meta":"","header-main-stick-meta":"","header-below-stick-meta":"","astra-migrate-meta-layouts":"set","ast-page-background-enabled":"default","ast-page-background-meta":{"desktop":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"ast-content-background-meta":{"desktop":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"footnotes":""},"categories":[8],"tags":[39,86,66],"class_list":["post-5596","post","type-post","status-publish","format-standard","hentry","category-tech","tag-ai","tag-llm","tag-google"],"rttpg_featured_image_url":null,"rttpg_author":{"display_name":"NullThought","author_link":"https:\/\/nullthought.net\/?author=1"},"rttpg_comment":0,"rttpg_category":"<a href=\"https:\/\/nullthought.net\/?cat=8\" rel=\"category\">Tech<\/a>","rttpg_excerpt":"\u8bba\u6587Evolving Deeper LLM Thinking\u63d0\u51fa\u4e86\u4e00\u79cd\u521b\u65b0\u7684\u8fdb\u5316\u641c\u7d22\u7b56\u7565\u2014\u2014Mind Evol&hellip;","_links":{"self":[{"href":"https:\/\/nullthought.net\/index.php?rest_route=\/wp\/v2\/posts\/5596","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/nullthought.net\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/nullthought.net\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/nullthought.net\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/nullthought.net\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=5596"}],"version-history":[{"count":2,"href":"https:\/\/nullthought.net\/index.php?rest_route=\/wp\/v2\/posts\/5596\/revisions"}],"predecessor-version":[{"id":5599,"href":"https:\/\/nullthought.net\/index.php?rest_route=\/wp\/v2\/posts\/5596\/revisions\/5599"}],"wp:attachment":[{"href":"https:\/\/nullthought.net\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=5596"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/nullthought.net\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=5596"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/nullthought.net\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=5596"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}