{"id":25378,"date":"2024-12-20T21:07:01","date_gmt":"2024-12-20T13:07:01","guid":{"rendered":"https:\/\/www.1ai.net\/?p=25378"},"modified":"2024-12-20T21:07:01","modified_gmt":"2024-12-20T13:07:01","slug":"%e7%be%8e%e5%9b%bd%e5%93%a5%e4%bc%a6%e6%af%94%e4%ba%9a%e5%a4%a7%e5%ad%a6%e7%a0%94%e7%a9%b6%ef%bc%9a%e5%a4%a7%e8%af%ad%e8%a8%80%e6%a8%a1%e5%9e%8b%e6%ad%a3%e5%9c%a8%e5%8f%98%e5%be%97%e8%b6%8a%e6%9d%a5","status":"publish","type":"post","link":"https:\/\/www.1ai.net\/en\/25378.html","title":{"rendered":"Columbia University Study: Big Language Models Are Becoming More Like the Human Brain"},"content":{"rendered":"<p>12 \u6708 20 \u65e5\u6d88\u606f\uff0c\u4ee5\u00a0ChatGPT \u7b49\u4e3a\u4ee3\u8868\u7684\u5927\u578b\u8bed\u8a00\u6a21\u578b\uff08LLMs\uff09\u5728\u8fc7\u53bb\u51e0\u5e74\u4e2d\u8d8a\u6765\u8d8a\u64c5\u957f\u5904\u7406\u548c\u751f\u6210\u4eba\u7c7b\u8bed\u8a00\uff0c\u4f46\u8fd9\u4e9b\u6a21\u578b\u5728\u591a\u5927\u7a0b\u5ea6\u4e0a\u6a21\u62df\u4e86\u4eba\u7c7b\u5927\u8111\u652f\u6301\u8bed\u8a00\u5904\u7406\u7684\u795e\u7ecf\u8fc7\u7a0b\uff0c\u8fd8\u6709\u5f85\u8fdb\u4e00\u6b65\u9610\u660e\u3002<\/p>\n<p>\u636e Tech Xplore 18 \u65e5\u62a5\u9053\uff0c<a href=\"https:\/\/www.1ai.net\/en\/tag\/%e5%93%a5%e4%bc%a6%e6%af%94%e4%ba%9a%e5%a4%a7%e5%ad%a6\" title=\"\u3010\u67e5\u770b\u542b\u6709[\u54e5\u4f26\u6bd4\u4e9a\u5927\u5b66]\u6807\u7b7e\u7684\u6587\u7ae0\u3011\" target=\"_blank\" rel=\"noopener\">\u54e5\u4f26\u6bd4\u4e9a\u5927\u5b66<\/a>\u548c\u8d39\u56e0\u65af\u5766\u533b\u5b66\u7814\u7a76\u6240\u7684\u7814\u7a76\u56e2\u961f\u8fd1\u671f\u8fdb\u884c\u4e86\u4e00\u9879\u7814\u7a76\uff0c\u63a2\u7d22\u4e86 LLM \u4e0e\u5927\u8111\u795e\u7ecf\u53cd\u5e94\u7684\u76f8\u4f3c\u6027\u3002\u7814\u7a76\u8868\u660e\uff0c\u968f\u7740 LLM \u6280\u672f\u7684\u8fdb\u6b65\uff0c\u8fd9\u4e9b\u6a21\u578b<strong>\u4e0d\u4ec5\u5728\u6027\u80fd\u4e0a\u6709\u6240\u63d0\u5347<\/strong>\uff0c\u800c\u4e14\u7ed3\u6784\u4e0a\u4e5f<strong>\u8d8a\u6765\u8d8a\u63a5\u8fd1\u4eba\u7c7b\u5927\u8111<\/strong>\u3002<\/p>\n<p>\u8bba\u6587\u7684\u7b2c\u4e00\u4f5c\u8005\u52a0\u6587\u30fb\u7c73\u65bd\u52d2\uff08Gavin Mischler\uff09\u5728\u63a5\u53d7 Tech Xplore \u91c7\u8bbf\u65f6\u8868\u793a\uff1a\u201c\u6211\u4eec\u64b0\u5199\u8fd9\u7bc7\u8bba\u6587\u7684\u7075\u611f\u6765\u6e90\u4e8e\u8fd1\u5e74\u6765 LLM \u548c\u795e\u7ecf AI \u7814\u7a76\u7684\u8fc5\u901f\u53d1\u5c55\u3002\u201d<\/p>\n<p>\u201c\u51e0\u5e74\u524d\uff0c\u6709\u51e0\u7bc7\u6587\u7ae0\u663e\u793a\uff0cGPT-2 \u7684\u8bcd\u5d4c\u5165\u4e0e\u4eba\u8111\u5bf9\u8bed\u8a00\u7684\u795e\u7ecf\u53cd\u5e94\u6709\u4e00\u5b9a\u76f8\u4f3c\u6027\uff0c\u4f46\u5728\u4eba\u5de5\u667a\u80fd\u8fd9\u4e00\u98de\u901f\u53d1\u5c55\u7684\u9886\u57df\u4e2d\uff0cGPT-2 \u5982\u4eca\u5df2\u7ecf\u8fc7\u65f6\uff0c\u7b97\u4e0d\u4e0a\u6700\u5f3a\u5927\u3002\u201d<\/p>\n<p>\u201c\u81ea\u4ece ChatGPT \u53d1\u5e03\u540e\uff0c\u6d8c\u73b0\u4e86\u8bb8\u591a\u66f4\u5f3a\u5927\u7684\u6a21\u578b\uff0c\u4f46\u5173\u4e8e\u8fd9\u4e9b\u65b0\u6a21\u578b\u662f\u5426\u4ecd\u7136\u8868\u73b0\u51fa\u76f8\u540c\u7684\u5927\u8111\u76f8\u4f3c\u6027\uff0c\u76f8\u5173\u7814\u7a76\u5374\u5e76\u4e0d\u591a\u3002\u201d<\/p>\n<p>\u7c73\u65bd\u52d2\u53ca\u5176\u56e2\u961f\u7684\u4e3b\u8981\u76ee\u6807\uff0c\u662f\u63a2\u7a76<strong>\u6700\u65b0\u4e00\u4ee3\u7684 LLM \u662f\u5426\u4f9d\u7136\u8868\u73b0\u51fa\u4e0e\u4eba\u7c7b\u5927\u8111\u76f8\u4f3c\u7684\u7279\u5f81<\/strong>\u3002\u7814\u7a76\u4eba\u5458\u5bf9 12 \u4e2a\u4e0d\u540c\u7684\u5f00\u6e90 LLM \u8fdb\u884c\u4e86\u5206\u6790\uff0c\u8fd9\u4e9b\u6a21\u578b\u5728\u67b6\u6784\u548c\u53c2\u6570\u6570\u91cf\u4e0a\u51e0\u4e4e\u4e00\u81f4\u3002\u540c\u65f6\uff0c\u4ed6\u4eec\u4e5f\u901a\u8fc7\u5728\u795e\u7ecf\u5916\u79d1\u60a3\u8005\u7684\u8111\u90e8\u690d\u5165\u7535\u6781\u8bb0\u5f55\u5176\u542c\u5230\u8bed\u8a00\u65f6\u7684\u5927\u8111\u53cd\u5e94\u3002<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-25379\" title=\"34e7eeb2j00sosn1x006td000m800dcp\" src=\"https:\/\/www.1ai.net\/wp-content\/uploads\/2024\/12\/34e7eeb2j00sosn1x006td000m800dcp.jpg\" alt=\"34e7eeb2j00sosn1x006td000m800dcp\" width=\"800\" height=\"480\" \/><\/p>\n<p>\u7c73\u65bd\u52d2\u89e3\u91ca\u9053\uff1a\u201c\u6211\u4eec\u8fd8\u5c06\u540c\u6837\u7684\u6f14\u8bb2\u6587\u672c\u8f93\u5165 LLM\uff0c\u5e76\u63d0\u53d6\u5176\u8bcd\u5d4c\u5165\uff0c\u8fd9\u4e9b\u5d4c\u5165\u662f\u6a21\u578b\u5185\u90e8\u7528\u6765\u5904\u7406\u548c\u7f16\u7801\u6587\u672c\u7684\u8868\u793a\u3002\u4e3a\u4e86\u8861\u91cf\u8fd9\u4e9b LLM \u4e0e\u5927\u8111\u7684\u76f8\u4f3c\u6027\uff0c\u6211\u4eec\u5c1d\u8bd5\u901a\u8fc7\u9884\u6d4b\u5927\u8111\u5bf9\u8bcd\u8bed\u53cd\u5e94\u7684\u795e\u7ecf\u6d3b\u52a8\u6765\u8bc4\u4f30\u5b83\u4eec\u7684\u5bf9\u5e94\u6027\u3002\u901a\u8fc7\u8fd9\u79cd\u65b9\u5f0f\uff0c\u6211\u4eec\u53ef\u4ee5\u4e86\u89e3\u4e24\u8005\u7684\u76f8\u4f3c\u5ea6\u3002\u201d<\/p>\n<p>\u5728\u6570\u636e\u6536\u96c6\u540e\uff0c\u7814\u7a76\u4eba\u5458\u5229\u7528\u8ba1\u7b97\u5de5\u5177\u5206\u6790 LLM \u4e0e\u5927\u8111\u7684\u5bf9\u9f50\u7a0b\u5ea6\u3002\u4ed6\u4eec\u7279\u522b\u5173\u6ce8\u54ea\u4e9b\u5c42\u6b21\u7684 LLM \u4e0e\u5927\u8111\u4e2d\u4e0e\u8bed\u8a00\u5904\u7406\u76f8\u5173\u7684\u533a\u57df\u6700\u4e3a\u5339\u914d\u3002\u5927\u8111\u5bf9\u8bed\u8a00\u7684\u53cd\u5e94\u5df2\u77e5\u4f1a\u9010\u6b65\u901a\u8fc7\u5bf9\u8bed\u97f3\u7684\u58f0\u5b66\u3001\u8bed\u97f3\u5b66\u7b49\u6210\u5206\u7684\u5206\u6790\uff0c\u5efa\u7acb\u8bed\u8a00\u7684\u62bd\u8c61\u8868\u5f81\u3002<\/p>\n<p>\u7c73\u65bd\u52d2\u8868\u793a\uff1a\u201c\u6211\u4eec\u53d1\u73b0\uff0c\u968f\u7740 LLM \u80fd\u529b\u7684\u63d0\u5347\uff0c\u8fd9\u4e9b\u6a21\u578b\u7684\u8bcd\u5d4c\u5165\u4e0e\u5927\u8111\u5bf9\u8bed\u8a00\u7684\u53cd\u5e94\u8d8a\u6765\u8d8a\u63a5\u8fd1\u3002\u66f4\u51fa\u4e4e\u610f\u6599\u7684\u662f\uff0c<strong>\u968f\u7740\u6a21\u578b\u6027\u80fd\u7684\u63d0\u5347\uff0c\u5b83\u4eec\u4e0e\u5927\u8111\u5c42\u6b21\u7ed3\u6784\u7684\u5bf9\u9f50\u7a0b\u5ea6\u4e5f\u6709\u6240\u63d0\u9ad8<\/strong>\u3002\u8fd9\u610f\u5473\u7740\uff0c\u8bed\u8a00\u5904\u7406\u8fc7\u7a0b\u4e2d\u5927\u8111\u4e0d\u540c\u533a\u57df\u63d0\u53d6\u7684\u4fe1\u606f\uff0c\u4e0e\u6027\u80fd\u8f83\u5f3a\u7684 LLM \u7684\u4e0d\u540c\u5c42\u6b21\u63d0\u53d6\u7684\u4fe1\u606f\u5bf9\u6bd4\uff0c\u66f4\u52a0\u4e00\u81f4\u3002\u201d<\/p>\n<p>\u8fd9\u4e9b\u7814\u7a76\u7ed3\u679c\u8868\u660e\uff0c\u8868\u73b0\u6700\u597d\u7684 LLM \u66f4\u80fd\u51c6\u786e\u53cd\u6620\u5927\u8111\u7684\u8bed\u8a00\u5904\u7406\u53cd\u5e94\u3002\u800c\u4e14\uff0c\u8fd9\u4e9b\u6a21\u578b\u7684\u4f18\u79c0\u8868\u73b0\u53ef\u80fd\u4e0e\u5176\u65e9\u671f\u5c42\u6b21\u7684\u9ad8\u6548\u6027\u6709\u5173\u3002<\/p>\n<p>\u6709\u5173\u7814\u7a76\u7ed3\u679c\u5df2\u7ecf\u5728\u300a\u81ea\u7136\u30fb\u673a\u5668\u667a\u80fd\u300b\u671f\u520a\u4e0a\u53d1\u8868\uff0c1AI\u9644\u94fe\u63a5\uff1a<a href=\"https:\/\/www.nature.com\/articles\/s42256-024-00925-4\">\u70b9\u6b64\u524d\u5f80<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Dec. 20, 2011 - Large language models (LLMs), represented by ChatGPT and others, have become increasingly adept at processing and generating human language over the past few years, but the extent to which these models mimic the neural processes of the human brain that support language processing remains to be further elucidated. As reported by Tech Xplore 18, a team of researchers from Columbia University and the Feinstein Institute for Medical Research recently conducted a study that explored the similarities between LLMs and neural responses in the brain. The study showed that as LLM technology advances, these models not only improve in performance, but also become structurally closer to the human brain. Gavin Mischler, the paper's first author, spoke with Tech<\/p>","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[146],"tags":[5120,706,236],"collection":[],"class_list":["post-25378","post","type-post","status-publish","format-standard","hentry","category-news","tag-5120","tag-706","tag-236"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/25378","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/comments?post=25378"}],"version-history":[{"count":0,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/25378\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/media?parent=25378"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/categories?post=25378"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/tags?post=25378"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/collection?post=25378"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}