{"id":12960,"date":"2025-07-19T12:19:57","date_gmt":"2025-07-19T04:19:57","guid":{"rendered":"https:\/\/ihower.tw\/blog\/?p=12960"},"modified":"2025-07-21T18:25:31","modified_gmt":"2025-07-21T10:25:31","slug":"ai-evals-and-error-analysis","status":"publish","type":"post","link":"https:\/\/ihower.tw\/blog\/12960-ai-evals-and-error-analysis","title":{"rendered":"\u4ec0\u9ebc\u662f AI \u61c9\u7528\u8a55\u4f30\u7684\u932f\u8aa4\u5206\u6790 Error Analysis?"},"content":{"rendered":"\n<p>\u6700\u8fd1\u5728\u4e0a <a href=\"https:\/\/maven.com\/parlance-labs\/evals\">Hamel + Shreya \u7684 AI Evals For Engineers &amp; PMs \u8ab2\u7a0b<\/a>\uff0c\u9019\u61c9\u8a72\u662f\u5e02\u9762\u4e0a\u6700\u6df1\u5165\u63a2\u8a0e AI \u61c9\u7528\u8a55\u4f30\u7684\u8ab2\u7a0b\u4e86\u3002\u4ee5\u4e0b\u6839\u64da\u7db2\u4e0a\u6709\u516c\u958b\u7684\u5167\u5bb9\uff0c\u6574\u7406\u4e00\u7bc7\u7cbe\u83ef\u5167\u5bb9(\u5927\u7d04\u662f\u8ab2\u7a0b\u7684\u524d1\/4\u5167\u5bb9)\u3002<\/p>\n\n\n\n<p>\u5982\u679c\u4f60\u6b63\u5728\u958b\u767c AI \u61c9\u7528\uff0c\u61c9\u8a72\u90fd\u9047\u904e\u9019\u7a2e\u60c5\u6cc1\uff1a\u7522\u54c1\u505a\u51fa\u4f86\u4e86\uff0c\u770b\u8d77\u4f86\u9084\u884c\uff0c\u4f46\u7e3d\u89ba\u5f97\u54ea\u88e1\u602a\u602a\u7684\u3002\u4f7f\u7528\u8005\u62b1\u6028\u4e00\u4e9b\u5947\u602a\u7684\u554f\u984c\uff0c\u4f46\u4f60\u4e0d\u77e5\u9053\u5f9e\u4f55\u6539\u8d77\u3002\u9019\u7bc7\u6587\u7ae0\u8981\u4ecb\u7d39\u7684\u5c31\u662f\u8a55\u4f30\u548c\u932f\u8aa4\u5206\u6790 Error Analysis \u7cfb\u7d71\u6027\u65b9\u6cd5\u3002<\/p>\n\n\n\n<!--more-->\n\n\n\n<h2 class=\"wp-block-heading\">\u70ba\u4ec0\u9ebc LLM \u8a55\u4f30\u9019\u9ebc\u96e3\uff1f<\/h2>\n\n\n\n<p>\u50b3\u7d71\u7684\u6a5f\u5668\u5b78\u7fd2\u8a55\u4f30\u65b9\u6cd5\u5728 LLM \u6642\u4ee3\u9762\u81e8\u5168\u65b0\u6311\u6230\uff1a<\/p>\n\n\n\n<ol class=\"wp-block-list\">\n<li><strong>\u8f38\u51fa\u7684\u975e\u7d50\u69cb\u5316\u7279\u6027<\/strong>\uff1a\u4e0d\u50cf\u50b3\u7d71 ML \u6709\u660e\u78ba\u7684\u6578\u503c\u6307\u6a19\uff08\u5982\u6e96\u78ba\u7387\u3001\u53ec\u56de\u7387\uff09\uff0cLLM \u8f38\u51fa\u7684\u662f\u958b\u653e\u5f0f\u6587\u672c\uff0c\u96e3\u4ee5\u7528\u7c21\u55ae\u7684\u6307\u6a19\u8861\u91cf<\/li>\n\n\n\n<li><strong>\u4e3b\u89c0\u6027\u554f\u984c<\/strong>\uff1a\u4ec0\u9ebc\u662f\u300c\u597d\u7684\u300d\u56de\u61c9\u5f80\u5f80\u9ad8\u5ea6\u4f9d\u8cf4\u5177\u9ad4\u60c5\u5883\u548c\u4f7f\u7528\u8005\u671f\u671b<\/li>\n\n\n\n<li><strong>\u908a\u7de3\u6848\u4f8b\u5c64\u51fa\u4e0d\u7aae<\/strong>\uff1a\u4f60\u6e2c\u8a66\u4e86 100 \u7a2e\u60c5\u6cc1\uff0c\u4f7f\u7528\u8005\u504f\u504f\u6703\u7528\u7b2c 101 \u7a2e\u65b9\u5f0f\u4f86\u554f<\/li>\n\n\n\n<li><strong>\u6301\u7e8c\u5c0d\u8a71\u7684\u8907\u96dc\u6027<\/strong>\uff1a\u4e0d\u50cf\u4e00\u554f\u4e00\u7b54\u5c31\u7d50\u675f\uff0c\u771f\u5be6\u5c0d\u8a71\u53ef\u80fd\u6301\u7e8c\u5e7e\u5929\u751a\u81f3\u5e7e\u500b\u6708<\/li>\n<\/ol>\n\n\n\n<h2 class=\"wp-block-heading\">\u4e09\u5927\u9d3b\u6e9d\u96e3\u95dc<\/h2>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"638\" data-attachment-id=\"12964\" data-permalink=\"https:\/\/ihower.tw\/blog\/12960-ai-evals-and-error-analysis\/image-20\" data-orig-file=\"https:\/\/ihower.tw\/blog\/wp-content\/uploads\/2025\/07\/image-4.png\" data-orig-size=\"2492,1552\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"image\" data-image-description=\"\" data-image-caption=\"\" data-medium-file=\"https:\/\/ihower.tw\/blog\/wp-content\/uploads\/2025\/07\/image-4-300x187.png\" data-large-file=\"https:\/\/ihower.tw\/blog\/wp-content\/uploads\/2025\/07\/image-4-1024x638.png\" src=\"https:\/\/ihower.tw\/blog\/wp-content\/uploads\/2025\/07\/image-4-1024x638.png\" alt=\"\" class=\"wp-image-12964\" srcset=\"https:\/\/ihower.tw\/blog\/wp-content\/uploads\/2025\/07\/image-4-1024x638.png 1024w, https:\/\/ihower.tw\/blog\/wp-content\/uploads\/2025\/07\/image-4-300x187.png 300w, https:\/\/ihower.tw\/blog\/wp-content\/uploads\/2025\/07\/image-4-768x478.png 768w, https:\/\/ihower.tw\/blog\/wp-content\/uploads\/2025\/07\/image-4-1536x957.png 1536w, https:\/\/ihower.tw\/blog\/wp-content\/uploads\/2025\/07\/image-4-2048x1275.png 2048w, https:\/\/ihower.tw\/blog\/wp-content\/uploads\/2025\/07\/image-4-1568x977.png 1568w\" sizes=\"auto, (max-width: 1024px) 100vw, 1024px\" \/><\/figure>\n\n\n\n<p>Shreya Shankar \u63d0\u51fa\u4e86\u4e00\u500b\u5f88\u6709\u6d1e\u5bdf\u529b\u7684\u6bd4\u55bb\uff0c\u958b\u767c AI \u61c9\u7528\u5c31\u50cf\u8981\u8de8\u8d8a\u4e09\u500b\u5927\u9d3b\u6e9d\uff1a<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">1. \u7406\u89e3\u9d3b\u6e9d: \u641e\u61c2\u5230\u5e95\u767c\u751f\u4ec0\u9ebc\u4e8b<\/h3>\n\n\n\n<p>\u7406\u89e3\u9d3b\u6e9d\u662f\u95dc\u65bc\u7406\u89e3\u6578\u64da\u3002\u7576\u6211\u5011\u8aaa\u300c\u770b\u6578\u64da\u300d\u6642\uff0c\u7522\u751f\u7684\u6210\u679c\u662f\u4f60\u8166\u4e2d\u5c0d\u6578\u64da\u5167\u5bb9\u548c\u5931\u6557\u6a21\u5f0f\u7684\u77e5\u8b58\u3002\u9019\u5c31\u50cf\u5075\u63a2\u8fa6\u6848\uff0c\u4f60\u9700\u8981\uff1a<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>\u771f\u6b63\u53bb\u770b\u4f7f\u7528\u8005\u5982\u4f55\u4f7f\u7528\u4f60\u7684\u7522\u54c1\uff0c\u7406\u89e3\u4ed6\u5011\u5be6\u969b\u5982\u4f55\u8207\u7cfb\u7d71\u4e92\u52d5<\/li>\n\n\n\n<li>\u89c0\u5bdf\u4eba\u5011\u6bcf\u5929\u5411\u4f60\u7684\u7522\u54c1\u63d0\u51fa\u4ec0\u9ebc\u554f\u984c\uff0c\u8b58\u5225\u4ed6\u5011\u60f3\u8981\u89e3\u6c7a\u7684\u75db\u9ede<\/li>\n\n\n\n<li>\u627e\u51fa AI \u5728\u54ea\u4e9b\u5730\u65b9\u6703\u51fa\u932f\uff0c\u8b58\u5225\u6578\u64da\u4e2d\u7684\u6a21\u5f0f\u548c\u7570\u5e38<\/li>\n\n\n\n<li>\u767c\u73fe\u96b1\u85cf\u7684\u5931\u6557\u6a21\u5f0f<\/li>\n\n\n\n<li>\u7406\u89e3\u4e0d\u540c\u6578\u64da\u7fa4\u96c6\u7684\u7279\u5fb5<\/li>\n\n\n\n<li>\u6316\u6398\u4f7f\u7528\u8005\u771f\u6b63\u7684\u9700\u6c42<\/li>\n<\/ul>\n\n\n\n<p>NurtureBoss \u7684 Jacob Carter \u5c0d\u6b64\u6df1\u6709\u9ad4\u6703\uff1a\u300c\u7576\u6211\u5011\u767c\u5e03\u7522\u54c1\u6642\uff0c\u5c0d\u65bc\u6b63\u5728\u767c\u751f\u7684\u4e8b\u60c5\u4f86\u8aaa\uff0c\u9019\u5c0d\u6211\u5011\u4f86\u8aaa\u662f\u4e00\u500b\u5de8\u5927\u7684\u9ed1\u76d2\u5b50\u3002\u300d\u4f46\u5728\u67e5\u770b\u4e86\u6578\u5343\u6b21\u5c0d\u8a71\u5f8c\uff0c\u4ed6\u767c\u73fe\uff1a\u300c\u80fd\u5920\u5be6\u969b\u770b\u5230\u4eba\u5011\u6bcf\u5929\u5411\u4f60\u7684\u7522\u54c1\u8a62\u554f\u4ec0\u9ebc\uff0c\u8b58\u5225\u51fa\u4ed6\u5011\u5c0b\u6c42\u89e3\u6c7a\u54ea\u4e9b\u554f\u984c\uff0c\u9019\u5c07\u70ba\u4f60\u5efa\u7acb\u7522\u54c1\u8def\u7dda\u5716\u3002\u300d<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">2. \u898f\u7bc4\u9d3b\u6e9d: \u628a\u8a71\u8aaa\u6e05\u695a<\/h3>\n\n\n\n<p>\u9019\u662f\u6700\u591a\u4eba\u683d\u8ddf\u982d\u7684\u5730\u65b9\u3002\u4f60\u4ee5\u70ba\u4f60\u7684\u6307\u4ee4\u5f88\u6e05\u695a\uff0c\u4f46 AI \u7684\u7406\u89e3\u53ef\u80fd\u5b8c\u5168\u4e0d\u4e00\u6a23\u3002\u5982\u4f55\u6210\u70ba\u4e00\u500b\u597d\u7684\u6e9d\u901a\u8005\uff1f\u56e0\u70ba\u6211\u5011\u5fc5\u9808\u4ee5\u6975\u5176\u8a73\u7d30\u7684\u65b9\u5f0f\u5411 LLM \u6307\u5b9a\u6240\u6709\u5167\u5bb9\uff0c\u4f7f\u5176\u7406\u89e3\u5fc5\u9808\u660e\u78ba\u7121\u6b67\u7fa9\uff0c\u9019\u975e\u5e38\u56f0\u96e3\u3002<\/p>\n\n\n\n<p>\u5e38\u898b\u7684\u932f\u8aa4\uff1a<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>\u53ea\u7d66\u4e00\u500b\u4f8b\u5b50<\/strong>\uff1a\u5f88\u591a\u4eba\u5b78\u6703 few-shot prompting \u5f8c\uff0c\u53ea\u7d66\u4e00\u500b\u4f8b\u5b50\u3002\u6a21\u578b\u53ef\u80fd\u6703\u904e\u5ea6\u95dc\u6ce8\u8a72\u8209\u4f8b\uff0c\u5c0e\u81f4\u6240\u6709\u56de\u7b54\u90fd\u9577\u5f97\u5f88\u50cf<\/li>\n\n\n\n<li><strong>\u6307\u4ee4\u904e\u5ea6\u898f\u7bc4<\/strong>\uff1a\u4f8b\u5982\u5beb\u300c\u8acb\u7576\u4e00\u500b\u503c\u5f97\u4fe1\u4efb\u7684\u9867\u554f\u300d\uff0c\u7d50\u679c AI \u8b8a\u5f97\u592a\u76f4\u767d\uff0c\u751a\u81f3\u958b\u59cb\u6279\u8a55\u4f7f\u7528\u8005\uff01<\/li>\n\n\n\n<li><strong>Prompt \u8d8a\u5beb\u8d8a\u9577<\/strong>\uff1a\u6700\u5f8c\u8b8a\u6210\u843d\u843d\u9577\u7684\u4f5c\u6587\u3002\u5176\u5be6\u5f88\u591a\u6642\u5019\uff0c\u522a\u6389\u4e00\u534a\u53cd\u800c\u66f4\u597d\u7528\u3002\u56e0\u70ba\u6211\u5011\u4e0d\u65b7\u5617\u8a66\u89e3\u6c7a\u898f\u7bc4\u9d3b\u6e9d\uff0c\u8a66\u5716\u8b93\u5b83\u8d8a\u4f86\u8d8a\u5177\u9ad4\uff0c\u5be6\u969b\u4e0a\u5fd8\u8a18\u4e86\u91cd\u69cb\u6211\u5011\u7684 prompts<\/li>\n<\/ul>\n\n\n\n<p>Jacob \u5206\u4eab\u4e86\u4e00\u500b AI \u6642\u9593\u611f\u6df7\u4e82\u7684\u666e\u904d\u554f\u984c\uff0c\u5ba2\u6236\u8aaa\u300e\u6211\u60f3\u5b89\u6392\u5169\u9031\u5f8c\u770b\u623f\u300f\uff0cAI \u6703\u56de\u7b54\u300e\u597d\u7684\uff0c\u90a3\u662f2\u670829\u65e5\u300f\u4f46\u6839\u672c\u4e0d\u662f\u5169\u9031\u5f8c\uff01\u6216\u8005\u6709\u4eba\u8aaa\u8981\u5b89\u63923\u67081\u65e5\u770b\u623f\uff0cAI \u537b\u53bb\u627e2020\u5e743\u67081\u65e5\u7684\u7a7a\u6a94\u3002<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">3. \u6cdb\u5316\u9d3b\u6e9d: \u78ba\u4fdd\u5230\u8655\u90fd\u80fd\u7528<\/h3>\n\n\n\n<p>\u9019\u662f\u50b3\u7d71\u7684\u6a5f\u5668\u5b78\u7fd2\u554f\u984c\uff1a\u300c\u6211\u5df2\u7d93\u5beb\u597d\u4e86 pipeline\uff0c\u4f46\u5b83\u80fd\u6cdb\u5316\u5230\u6211\u7684\u6578\u64da\u55ce\uff1f\u5b83\u80fd\u6cdb\u5316\u5230\u5176\u4ed6\u4eba\u7684\u6578\u64da\u55ce\uff1f\u5982\u679c\u6211\u5728\u5be6\u969b\u74b0\u5883\u4e2d\u90e8\u7f72\u5b83\uff0c\u5b83\u6703\u5de5\u4f5c\u55ce\uff1f\u300d<\/p>\n\n\n\n<p>\u5728\u4f60\u7684\u6e2c\u8a66\u74b0\u5883\u8dd1\u5f97\u597d\u597d\u7684\uff0c\u4e00\u4e0a\u7dda\u5c31\u5404\u7a2e\u51fa\u5305\u3002\u7e3d\u662f\u6709\u4e9b\u908a\u908a\u89d2\u89d2\u7684\u60c5\u6cc1\u4f60\u6c92\u60f3\u5230\u3002\u5728\u5be6\u969b\u4f7f\u7528\u4e2d\uff0c\u9019\u4e9b\u5947\u602a\u7684\u60c5\u6cc1\u4e00\u5b9a\u6703\u51fa\u73fe\u3002<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">\u4ec0\u9ebc\u662f\u6539\u9032\u5faa\u74b0? \u5206\u6790\u3001\u6e2c\u91cf\u3001\u6539\u9032<\/h2>\n\n\n\n<p>LLM \u61c9\u7528\u6539\u9032\u7684\u6838\u5fc3\u662f\u4e00\u500b\u4e09\u6b65\u9a5f\u5faa\u74b0\uff1a<\/p>\n\n\n\n<ol class=\"wp-block-list\">\n<li><strong>\u5206\u6790\uff08Analyze\uff09<\/strong>\uff1a\u9032\u884c\u932f\u8aa4\u5206\u6790<\/li>\n\n\n\n<li><strong>\u6e2c\u91cf\uff08Measure\uff09<\/strong>\uff1a\u5927\u898f\u6a21\u6e2c\u91cf<\/li>\n\n\n\n<li><strong>\u6539\u9032\uff08Improve\uff09<\/strong>\uff1a\u57fa\u65bc\u6e2c\u91cf\u7d50\u679c\u9032\u884c\u6539\u9032<\/li>\n<\/ol>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"844\" data-attachment-id=\"12965\" data-permalink=\"https:\/\/ihower.tw\/blog\/12960-ai-evals-and-error-analysis\/image-21\" data-orig-file=\"https:\/\/ihower.tw\/blog\/wp-content\/uploads\/2025\/07\/image-5.png\" data-orig-size=\"2072,1708\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"image\" data-image-description=\"\" data-image-caption=\"\" data-medium-file=\"https:\/\/ihower.tw\/blog\/wp-content\/uploads\/2025\/07\/image-5-300x247.png\" data-large-file=\"https:\/\/ihower.tw\/blog\/wp-content\/uploads\/2025\/07\/image-5-1024x844.png\" src=\"https:\/\/ihower.tw\/blog\/wp-content\/uploads\/2025\/07\/image-5-1024x844.png\" alt=\"\" class=\"wp-image-12965\" srcset=\"https:\/\/ihower.tw\/blog\/wp-content\/uploads\/2025\/07\/image-5-1024x844.png 1024w, https:\/\/ihower.tw\/blog\/wp-content\/uploads\/2025\/07\/image-5-300x247.png 300w, https:\/\/ihower.tw\/blog\/wp-content\/uploads\/2025\/07\/image-5-768x633.png 768w, https:\/\/ihower.tw\/blog\/wp-content\/uploads\/2025\/07\/image-5-1536x1266.png 1536w, https:\/\/ihower.tw\/blog\/wp-content\/uploads\/2025\/07\/image-5-2048x1688.png 2048w, https:\/\/ihower.tw\/blog\/wp-content\/uploads\/2025\/07\/image-5-1568x1293.png 1568w\" sizes=\"auto, (max-width: 1024px) 100vw, 1024px\" \/><\/figure>\n\n\n\n<p><\/p>\n\n\n\n<h2 class=\"wp-block-heading\">\u4ec0\u9ebc\u662f\u932f\u8aa4\u5206\u6790 Error Analysis ?<\/h2>\n\n\n\n<p>\u7b2c\u4e00\u6b65\u7684\u932f\u8aa4\u5206\u6790\u662f\u6574\u500b\u6d41\u7a0b\u7684\u95dc\u9375\u3002\u91cd\u9ede\u5c31\u662f\u900f\u904e\u5206\u6790\u627e\u5230\u6709\u54ea\u4e9b\u5931\u6557\u6a21\u5f0f\uff0c\u5f8c\u7e8c\u6b65\u9a5f\u76f4\u63a5\u91dd\u5c0d\u9019\u4e9b\u5931\u6557\u6a21\u5f0f\u4f86\u9032\u884c\u91cf\u6e2c\u53ca\u6539\u9032\u3002<\/p>\n\n\n\n<blockquote class=\"wp-block-quote is-layout-flow wp-block-quote-is-layout-flow\">\n<p>\u7de8\u6309: \u91dd\u5c0d\u6c92\u6709\u6a19\u6e96\u7b54\u6848\u7684\u554f\u7b54\u8a55\u4f30(\u5c0d\u6bd4\u6709\u6a19\u6e96\u7b54\u6848\u7684\u662f\u6307\u55ae\u9078\u3001\u591a\u9078\u7b49\u6709\u56fa\u5b9a\u7b54\u6848)\uff0c\u9019\u88cf\u4e0d\u540c\u65bc\u5e38\u898b\u7684 <a href=\"https:\/\/arxiv.org\/abs\/2303.16634\">G-Eval <\/a>\u8a55\u4f30\u65b9\u5f0f\u63a1\u7528\u6b63\u9762\u8868\u5217\uff0c\u6839\u64da\u4f60\u7684 Criteria \u505a\u8a55\u4f30\u91cf\u6e2c\u6253\u5206(\u4f8b\u59821~5\u5206\u6709\u591a\u7b26\u5408)\u3002\u9019\u88cf\u6559\u7684\u65b9\u6cd5\u662f\u5148\u505a\u932f\u8aa4\u5206\u6790\uff0c\u62ff\u5230\u5177\u9ad4\u7684\u8ca0\u9762\u8868\u5217\u5f8c\uff0c\u5f8c\u7e8c\u518d\u91dd\u5c0d &#8220;\u6bcf\u4e00\u7a2e&#8221; \u5931\u6557\u6a21\u5f0f\u90fd\u4f86\u505a\u8a55\u4f30\u91cf\u6e2c\u548c\u6539\u9032\u3002<\/p>\n<\/blockquote>\n\n\n\n<p><\/p>\n\n\n\n<figure class=\"wp-block-image size-full is-resized\"><img loading=\"lazy\" decoding=\"async\" width=\"854\" height=\"708\" data-attachment-id=\"12961\" data-permalink=\"https:\/\/ihower.tw\/blog\/12960-ai-evals-and-error-analysis\/image-19\" data-orig-file=\"https:\/\/ihower.tw\/blog\/wp-content\/uploads\/2025\/07\/image-3.png\" data-orig-size=\"854,708\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"image\" data-image-description=\"\" data-image-caption=\"\" data-medium-file=\"https:\/\/ihower.tw\/blog\/wp-content\/uploads\/2025\/07\/image-3-300x249.png\" data-large-file=\"https:\/\/ihower.tw\/blog\/wp-content\/uploads\/2025\/07\/image-3.png\" src=\"https:\/\/ihower.tw\/blog\/wp-content\/uploads\/2025\/07\/image-3.png\" alt=\"\" class=\"wp-image-12961\" style=\"width:761px;height:auto\" srcset=\"https:\/\/ihower.tw\/blog\/wp-content\/uploads\/2025\/07\/image-3.png 854w, https:\/\/ihower.tw\/blog\/wp-content\/uploads\/2025\/07\/image-3-300x249.png 300w, https:\/\/ihower.tw\/blog\/wp-content\/uploads\/2025\/07\/image-3-768x637.png 768w\" sizes=\"auto, (max-width: 854px) 100vw, 854px\" \/><\/figure>\n\n\n\n<h3 class=\"wp-block-heading\">1. \u5275\u5efa\u521d\u59cb\u6578\u64da\u96c6\uff1a100 \u500b\u6a23\u672c<\/h3>\n\n\n\n<p>\u76ee\u6a19\u662f\u5f97\u5230 100 \u500b\u8de8\u8d8a\u4e0d\u540c\u4f7f\u7528\u7dad\u5ea6\u7684\u8f38\u5165\u3002\u70ba\u4ec0\u9ebc\u662f 100 \u500b\uff1f\u300c\u6c92\u4ec0\u9ebc\u7279\u5225\u539f\u56e0\uff0c\u5c31\u662f\u500b\u4e0d\u591a\u4e0d\u5c11\u7684\u6578\u5b57\uff0c\u8db3\u5920\u8b93\u4f60\u958b\u59cb\u4e86\u3002\u300d<\/p>\n\n\n\n<p>\u5177\u9ad4\u57f7\u884c\u6b65\u9a5f\uff1a<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>\u7dad\u5ea6\u601d\u8003<\/strong>\uff1a\u5728\u4f60\u7684\u61c9\u7528\u53ef\u80fd\u671f\u671b\u7d93\u6b77\u7684\u7dad\u5ea6\u9032\u884c\u63a1\u6a23\uff0c\u60f3\u51fa\u81f3\u5c11\u4e09\u500b\u7dad\u5ea6\u3002\u53ef\u4ee5\u5f9e\u529f\u80fd\u3001\u89d2\u8272\u3001\u67e5\u8a62\u8907\u96dc\u6027\u6216\u4f7f\u7528\u5834\u666f\u7684\u89d2\u5ea6\u4f86\u601d\u8003<\/li>\n\n\n\n<li><strong>\u7d44\u5408\u751f\u6210<\/strong>\uff1a\u751f\u6210\u9019\u4e09\u500b\u7dad\u5ea6\u7684 50 \u500b\u7d44\u5408\uff0c\u904e\u6ffe\u6389\u4e0d\u5408\u7406\u7684<\/li>\n\n\n\n<li><strong>\u67e5\u8a62\u751f\u6210<\/strong>\uff1a\u624b\u5beb\u6216\u4f7f\u7528 LLM \u5e6b\u52a9\u751f\u6210\u5b8c\u6574\u7684 100 \u500b\u73fe\u5be6\u67e5\u8a62<\/li>\n<\/ul>\n\n\n\n<h3 class=\"wp-block-heading\">2. \u6aa2\u8996\u8cc7\u6599<\/h3>\n\n\n\n<p>\u67e5\u770b\u8ffd\u8e64\u8a18\u9304\uff0c\u4e26\u5728\u8a18\u9304\u4e0a\u5beb\u7b46\u8a18\u3002\u67e5\u770b 100 \u500b\u6578\u64da\u9805\u76ee\u4e2d\u7684\u6bcf\u4e00\u500b\uff0c\u4e26\u5c0d\u4f60\u5728\u6578\u64da\u4e2d\u89c0\u5bdf\u5230\u7684\u5931\u6557\u6a21\u5f0f\u9032\u884c\u89c0\u5bdf\u3002<\/p>\n\n\n\n<p>\u95dc\u9375\u539f\u5247\uff1a<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>\u8b93\u985e\u5225\u5f9e\u6578\u64da\u4e2d\u81ea\u7136\u6d6e\u73fe\uff0c\u800c\u4e0d\u662f\u5e36\u8457\u9810\u8a2d\u60f3\u6cd5<\/li>\n\n\n\n<li>\u4e0d\u9700\u8981\u505a\u6839\u672c\u539f\u56e0\u5206\u6790(why \u767c\u751f)\uff0c\u53ea\u9700\u95dc\u6ce8\u89c0\u5bdf\u5230\u7684\u884c\u70ba\u548c\u6a21\u5f0f<\/li>\n\n\n\n<li>\u9810\u8a08\u82b1\u8cbb\u6642\u9593\uff1a\u300c\u9019\u662f\u4f60\u5c07\u82b1\u8cbb 80% \u6642\u9593\u7684\u5730\u65b9\uff0c\u5c0d\u65bc 100 \u500b\u8ffd\u8e64\u8a18\u9304\u53ef\u80fd\u9700\u8981\u5927\u7d04\u4e00\u500b\u5c0f\u6642\u300d<\/li>\n<\/ul>\n\n\n\n<p>\u53e6\u5916\uff0cNurtureBoss \u7684\u7d93\u9a57\u986f\u793a\uff0c\u5982\u679c\u81ea\u5df1\u958b\u767c\u7684\u6aa2\u8996\u5de5\u5177\u66f4\u80fd\u5927\u5e45\u63d0\u5347\u6548\u7387:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>\u80fd\u6e05\u695a\u986f\u793a\u5c0d\u8a71\u5168\u8c8c\uff1a\u7528\u6236\u8aaa\u4e86\u9019\u500b\uff0cAI \u8aaa\u4e86\u90a3\u500b\uff0cAI \u547c\u53eb\u4e86\u5de5\u5177\u3002\u9019\u662f AI \u5f9e\u5de5\u5177\u5f97\u5230\u7684\u56de\u61c9<\/li>\n\n\n\n<li>\u5feb\u901f\u6a19\u8a18\u548c\u8a3b\u89e3\uff1a\u6a19\u8a18\u9019\u500b\u5c0d\u8a71\u662f\u597d\u662f\u58de\uff0c\u7136\u5f8c\u5feb\u901f\u8f38\u5165\u8a3b\u89e3\u89e3\u91cb\u539f\u56e0<\/li>\n\n\n\n<li>\u5feb\u901f\u5206\u985e\u5931\u6557\u6a21\u5f0f\uff1a\u5206\u985e\u9019\u662f \u770b\u623f\u5b89\u6392\u932f\u8aa4\u3001\u672a\u89f8\u767c\u8f49\u63a5\u3001\u91cd\u8907\u8a62\u554f&#8230;\uff0c\u7136\u5f8c\u7cfb\u7d71\u7d71\u8a08\u54ea\u985e\u932f\u8aa4\u6700\u5e38\u767c\u751f<\/li>\n<\/ul>\n\n\n\n<h3 class=\"wp-block-heading\">3. \u5206\u7fa4\u6b78\u985e<\/h3>\n\n\n\n<p>\u770b\u5b8c\u6240\u6709\u6848\u4f8b\u5f8c\uff0c\u628a\u985e\u4f3c\u7684\u554f\u984c\u653e\u5728\u4e00\u8d77\u3002\u4f60\u53ef\u4ee5\u7528 AI \u5e6b\u5fd9\uff0c\u4f46\u6700\u5f8c\u4e00\u5b9a\u8981\u81ea\u5df1\u6aa2\u67e5\u3002Eugene \u8aaa\u300c\u6211\u628a\u6240\u6709\u6771\u897f\u90fd\u5148\u8b93 AI \u8dd1\u4e00\u904d\u7576\u8349\u7a3f\u3002AI \u6703\u7d66\u51fa\u4e0d\u932f\u7684\u5206\u985e\uff0c\u4f46\u6700\u5f8c\u7e3d\u662f\u9700\u8981 5-10% \u7684\u4eba\u5de5\u8abf\u6574\u3002\u300d<\/p>\n\n\n\n<p>\u901a\u904e\u5c07\u76f8\u4f3c\u7684\u5931\u6557\u6a21\u5f0f\u5206\u7d44\uff0c\u5efa\u69cb\u548c\u5408\u4f75\u51fa\u4f60\u61c9\u7528\u7684\u5931\u6557\u5206\u985e\u6cd5\uff0c\u7d71\u8a08\u6bcf\u7a2e\u932f\u8aa4\u51fa\u73fe\u7684\u983b\u7387\uff0c\u8b58\u5225\u6700\u5e38\u898b\u7684\u5931\u6557\u6a21\u5f0f\u3002<\/p>\n\n\n\n<p>\u5be6\u7528\u5efa\u8b70\uff1a<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>\u5617\u8a66\u5275\u5efa\u4e8c\u5143\u7684\u5931\u6557\u6a21\u5f0f\uff08\u53ef\u89c0\u5bdf\u5230\u7684 True \u6216 False \uff09\uff0c\u9019\u6a23\u66f4\u5bb9\u6613\u6709\u660e\u78ba\u7684\u5b9a\u7fa9\uff0c\u6bd4\u8f03\u7c21\u55ae<\/li>\n\n\n\n<li>\u59cb\u7d42\u624b\u52d5\u5be9\u67e5\u3001\u6539\u9032\u548c\u81ea\u884c\u5b9a\u7fa9\u9019\u4e9b\u5931\u6557\u6a21\u5f0f<\/li>\n<\/ul>\n\n\n\n<h3 class=\"wp-block-heading\">4. \u6a19\u8a18\u66f4\u591a\u8ffd\u8e64\u8a18\u9304\u4e26\u8fed\u4ee3<\/h3>\n\n\n\n<p>\u5728\u9019\u500b\u904e\u7a0b\u4e2d\uff0c\u4e0d\u8981\u64d4\u5fc3\u4f60\u7684\u5931\u6557\u6a21\u5f0f\u547d\u540d\u6216\u5b9a\u7fa9\u53ef\u80fd\u6703\u6f14\u8b8a\u3002\u9019\u662f\u6a19\u8a3b\u6578\u64da\u6642\u5e38\u898b\u7684\u73fe\u8c61\uff0c\u96a8\u8457\u4f60\u6aa2\u67e5\u65b0\u7684\u8f38\u51fa\uff0c\u6a19\u6e96\u6703\u6f02\u79fb\u3002\u5be6\u969b\u4e0a\u4f60\u61c9\u8a72\u66f4\u9ad8\u8208\uff0c\u56e0\u70ba\u9019\u53cd\u6620\u4e86\u4f60\u5c0d\u6578\u64da\u7406\u89e3\u7684\u52a0\u6df1\u3002<\/p>\n\n\n\n<p>\u90a3\u8981\u6301\u7e8c\u6a19\u8a3b\u591a\u5c11\u8cc7\u6599? \u7576\u65b0\u7684\u8cc7\u6599\u90fd\u6c92\u767c\u73fe\u65b0\u7684\u5931\u6557\u6a21\u5f0f\u6642\uff0c\u5c31\u53ef\u4ee5\u505c\u4e0b\u4f86\u8fed\u4ee3\u4e86\u3002<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">\u7e3d\u7d50<\/h2>\n\n\n\n<p><strong>\u5fc5\u9808\u4eba\u5de5\u67e5\u770b\u6578\u64da<\/strong>\uff0c\u5f15\u7528 Greg Brockman \u63a8\u6587:\u300c\u624b\u52d5\u6aa2\u67e5\u6578\u64da\u662f\u6a5f\u5668\u5b78\u7fd2\u4e2d\u50f9\u503c\u8207\u8072\u671b\u6bd4\u6700\u9ad8\u7684\u6d3b\u52d5\u300d\u3002Hamel \u66f4\u9032\u4e00\u6b65\u5f37\u8abf\uff1a\u300c\u6211\u6703\u8aaa\u9019\u662f\u6a5f\u5668\u5b78\u7fd2\u4e2d\u6295\u8cc7\u5831\u916c\u7387\u6700\u9ad8\u7684\u6d3b\u52d5\uff0c\u800c\u4e14\u662f\u5efa\u7acb\u4efb\u4f55 AI \u7522\u54c1\u6642\u6295\u8cc7\u5831\u916c\u7387\u6700\u9ad8\u7684\u6d3b\u52d5\u3002\u300d\u96d6\u7136\u67e5\u770b\u6578\u64da\u611f\u89ba\u50cf\u67af\u71e5\u7684\u82e6\u5de5\uff0c\u4f46\u5be6\u969b\u4e0a\u300c\u7576\u4f60\u67e5\u770b\u6578\u64da\u6642\uff0c\u6703\u975e\u5e38\u5feb\u901f\u5730\u7372\u5f97\u5927\u91cf\u50f9\u503c\u300d\u3002<\/p>\n\n\n\n<figure class=\"wp-block-embed is-type-rich is-provider-twitter wp-block-embed-twitter\"><div class=\"wp-block-embed__wrapper\">\n<div class=\"embed-twitter\"><blockquote class=\"twitter-tweet\" data-width=\"550\" data-dnt=\"true\"><p lang=\"en\" dir=\"ltr\">Manual inspection of data has probably the highest value-to-prestige ratio of any activity in machine learning.<\/p>&mdash; Greg Brockman (@gdb) <a href=\"https:\/\/twitter.com\/gdb\/status\/1622683988736479232?ref_src=twsrc%5Etfw\">February 6, 2023<\/a><\/blockquote><script async src=\"https:\/\/platform.twitter.com\/widgets.js\" charset=\"utf-8\"><\/script><\/div>\n<\/div><\/figure>\n\n\n\n<p><strong>\u4e0d\u8981\u8ff7\u4fe1\u901a\u7528\u7684\u8a55\u4f30\u6307\u6a19<\/strong>\uff0c\u4f8b\u5982\u7c21\u6f54\u6027\u8a55\u5206\u3001\u5e7b\u89ba\u8a55\u5206\u7b49\u901a\u7528\u6307\u6a19\uff0c\u90a3\u4e9b\u90fd\u662f\u7121\u610f\u7fa9\u7684\u5b78\u8853\u7df4\u7fd2\u3002Jacob \u7684\u7d93\u9a57\u8b49\u5be6\u4e86\u9019\u9ede\uff0c\u4ed6\u5011\u900f\u904e\u932f\u8aa4\u5206\u6790\uff0c\u5c07\u65e5\u671f\u8655\u7406\u7684\u6210\u529f\u7387\u5f9e 33% \u63d0\u5347\u5230 100%\uff0c\u9019\u7a2e\u5177\u9ad4\u3001\u53ef\u6e2c\u91cf\u7684\u6539\u9032\u624d\u662f\u771f\u6b63\u6709\u50f9\u503c\u7684\u3002 \u4f60\u7684\u7cfb\u7d71\u6709\u5176\u7368\u7279\u6027\uff0c\u901a\u7528\u5de5\u5177\u548c\u6307\u6a19\u5f80\u5f80\u7121\u6cd5\u5b8c\u7f8e\u9069\u7528\u3002\u5c08\u6ce8\u65bc\u4f60\u7684\u5be6\u969b\u75db\u9ede\uff0c\u800c\u4e0d\u662f\u8ffd\u6c42\u6f02\u4eae\u7684\u901a\u7528\u5206\u6578\u3002<\/p>\n\n\n\n<p><strong>\u4e0d\u8981\u4e00\u958b\u59cb\u5c31\u7528 LLM \u505a\u521d\u6b65\u6a19\u8a3b<\/strong>\uff0c\u9019\u662f\u5f88\u591a\u4eba\u6703\u72af\u7684\u932f\u8aa4\u3002\u4e0d\u8981\u8b93 LLM \u5e6b\u4f60\u505a\u521d\u6b65\u6a19\u8a3b\u518d\u8abf\u6574\uff0c\u56e0\u70ba LLM \u6703\u628a\u4f60\u5e36\u504f\uff0c\u4f60\u6703\u88ab\u5b83\u7684\u5224\u65b7\u5f71\u97ff\uff0c\u4f60\u9700\u8981\u89aa\u81ea\u770b\u6578\u64da\uff0c\u5efa\u7acb\u81ea\u5df1\u5c0d\u8cc7\u6599\u7684\u76f4\u89ba\u7406\u89e3\uff0cLLM \u6293\u4e0d\u5230\u4f60\u5728\u610f\u7684\u300cvibes\u300d\u6c1b\u570d! \u82b1\u4e00\u500b\u5c0f\u6642\u6a19\u8a3b\u7d55\u5c0d\u503c\u5f97\u3002\u4f60\u6b63\u5728\u57fa\u65bc\u9019\u8a55\u4f30\u4f86\u5efa\u7acb\u6574\u500b\u7522\u54c1\uff0c\u767e\u5206\u4e4b\u767e\u503c\u5f97\u89aa\u81ea\u67e5\u770b\u4f60\u7684\u6578\u64da\u3002<\/p>\n\n\n\n<p>LLM \u61c9\u7528\u7684\u8a55\u4f30\u548c\u932f\u8aa4\u5206\u6790\u4e0d\u50c5\u662f\u6280\u8853\u554f\u984c\uff0c\u66f4\u662f\u4e00\u7a2e\u601d\u7dad\u65b9\u5f0f\u7684\u8f49\u8b8a\u3002\u6210\u529f\u7684\u95dc\u9375\u4e0d\u5728\u65bc\u4f7f\u7528\u4ec0\u9ebc\u5de5\u5177\u6216\u6846\u67b6\uff0c\u800c\u5728\u65bc\u5efa\u7acb\u4e00\u500b\u6301\u7e8c\u5b78\u7fd2\u548c\u6539\u9032\u7684\u6587\u5316\u3002\u900f\u904e\u7cfb\u7d71\u5316\u7684\u65b9\u6cd5\u3001\u6b63\u78ba\u7684\u5fc3\u614b\u548c\u6301\u7e8c\u7684\u52aa\u529b\uff0c\u53ef\u4ee5\u5c07\u4f60\u7684 LLM \u61c9\u7528\u5f9e\u6982\u5ff5\u9a57\u8b49\u63a8\u9032\u5230\u771f\u6b63\u6539\u8b8a\u7528\u6236\u751f\u6d3b\u7684\u7522\u54c1\u3002<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">\u53c3\u8003\u548c\u63a8\u85a6\u5167\u5bb9<\/h2>\n\n\n\n<p>\u8ab2\u7a0b\u5b78\u54e1 Alex S. \u7684\u5fc3\u5f97\u6587\u7ae0:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><a href=\"https:\/\/mlops.systems\/posts\/2025-05-20-how-to-think-about-evals.html\">How to think about evals<\/a><\/li>\n\n\n\n<li><a href=\"https:\/\/mlops.systems\/posts\/2025-05-23-error-analysis-to-find-failure-modes.html\">Error analysis to find failure modes<\/a><\/li>\n<\/ul>\n\n\n\n<p>\u4e09\u5834 Hamel \u7684 Youtube \u9304\u5f71:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><a href=\"https:\/\/www.youtube.com\/watch?v=qH1dZ8JLLdU\">Intro To Error Analysis: Creating Custom Data Annotation Apps<\/a>\uff0c\u9019\u662f\u6211\u6574\u7406\u9010\u5b57\u7a3f\u5f8c\u7684<a href=\"https:\/\/claude.ai\/public\/artifacts\/1bc6ef93-366d-4845-b7a8-a93d855bcbf5\">\u6587\u7ae0<\/a><\/li>\n\n\n\n<li><a href=\"https:\/\/www.youtube.com\/watch?v=e2i6JbU2R-s\">Nurture Boss \u8a2a\u8ac7\u6848\u4f8b<\/a>\uff0c\u9019\u662f\u6211\u6574\u7406\u9010\u5b57\u7a3f\u5f8c\u7684<a href=\"https:\/\/claude.ai\/public\/artifacts\/99edc756-acd2-4664-b375-95ab9fe85078\">\u6587\u7ae0<\/a><\/li>\n\n\n\n<li><a href=\"https:\/\/www.youtube.com\/watch?v=OJItZndMUII\">Hamel + Shreya \u548c Eugene \u7684\u8a2a\u8ac7<\/a>\uff0c\u9019\u662f\u6211\u6574\u7406\u9010\u5b57\u7a3f\u5f8c\u7684<a href=\"https:\/\/claude.ai\/public\/artifacts\/6647dc2a-671d-474a-b784-71d1a1ccd093\">\u6587\u7ae0<\/a><\/li>\n<\/ul>\n\n\n\n<p>\u53e6\u5916\u63a8\u85a6 Hamel \u7684\u6587\u7ae0: <\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><a href=\"https:\/\/hamel.dev\/blog\/posts\/field-guide\/index.html\">A Field Guide to Rapidly Improving AI Products<\/a>\uff0c\u91cd\u9ede\u5305\u62ec:\n<ul class=\"wp-block-list\">\n<li>\u932f\u8aa4\u5206\u6790\u624d\u662f\u738b\u9053\uff0c\u5225\u6c88\u8ff7\u6f02\u4eae\u7684 dashboard \u901a\u7528\u6307\u6a19<\/li>\n\n\n\n<li>\u6700\u91cd\u8981\u7684\u6295\u8cc7\uff1a\u5ba2\u88fd\u5316\u7684\u6578\u64da\u6aa2\u8996\u4ecb\u9762<\/li>\n\n\n\n<li>\u8b93\u9818\u57df\u5c08\u5bb6\u76f4\u63a5\u5beb Prompt<\/li>\n\n\n\n<li>\u7528\u5408\u6210\u6578\u64da\u8d77\u6b65<\/li>\n\n\n\n<li>\u4fdd\u6301\u8a55\u4f30\u7cfb\u7d71\u7684\u53ef\u4fe1\u5ea6\uff0c\u7528\u4e8c\u5143\u5224\u65b7\u53d6\u4ee3\u6a21\u7cca\u5206\u6578<\/li>\n\n\n\n<li>\u8def\u7dda\u5716\u8981\u6578\u5be6\u9a57\uff0c\u4e0d\u662f\u6578\u529f\u80fd<\/li>\n<\/ul>\n<\/li>\n\n\n\n<li><a href=\"https:\/\/hamel.dev\/blog\/posts\/evals-faq\/index.html\">AI Evals \u8ab2\u7a0b FAQ<\/a>\uff0c\u91cd\u9ede\u5305\u62ec:\n<ul class=\"wp-block-list\">\n<li>\u932f\u8aa4\u5206\u6790 (Error Analysis) \u662f\u738b\u9053<\/li>\n\n\n\n<li>\u81ea\u5efa\u8a55\u4f30\u4ecb\u9762\u6bd4\u73fe\u6210\u5de5\u5177\u597d<\/li>\n\n\n\n<li>\u4e8c\u5143\u8a55\u4f30 &gt; \u674e\u514b\u7279\u91cf\u8868(1-5\u5206)<\/li>\n\n\n\n<li>RAG \u6c92\u6b7b\uff0c\u53ea\u662f\u8981\u7528\u5c0d\u65b9\u6cd5<\/li>\n\n\n\n<li>\u5225\u7528\u73fe\u6210\u7684\u901a\u7528\u6307\u6a19\uff0c\u9019\u4e9b\u6307\u6a19\u5c0d\u5927\u90e8\u5206 AI \u61c9\u7528\u90fd\u6c92\u7528<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n\n\n\n<p><\/p>\n","protected":false},"excerpt":{"rendered":"<p>\u6700\u8fd1\u5728\u4e0a Hamel + Shreya \u7684 AI Evals For Engineers &amp; PMs  &hellip; <\/p>\n<p class=\"link-more\"><a href=\"https:\/\/ihower.tw\/blog\/12960-ai-evals-and-error-analysis\" class=\"more-link\">\u95b1\u8b80\u5168\u6587<span class=\"screen-reader-text\">\u3008\u4ec0\u9ebc\u662f AI \u61c9\u7528\u8a55\u4f30\u7684\u932f\u8aa4\u5206\u6790 Error Analysis?\u3009<\/span><\/a><\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_monsterinsights_skip_tracking":false,"_monsterinsights_sitenote_active":false,"_monsterinsights_sitenote_note":"","_monsterinsights_sitenote_category":0,"jetpack_post_was_ever_published":false,"_jetpack_newsletter_access":"","_jetpack_dont_email_post_to_subs":false,"_jetpack_newsletter_tier_id":0,"_jetpack_memberships_contains_paywalled_content":false,"_jetpack_memberships_contains_paid_content":false,"footnotes":"","jetpack_publicize_message":"","jetpack_publicize_feature_enabled":true,"jetpack_social_post_already_shared":true,"jetpack_social_options":{"image_generator_settings":{"template":"highway","default_image_id":0,"font":"","enabled":false},"version":2}},"categories":[80],"tags":[],"class_list":["post-12960","post","type-post","status-publish","format-standard","hentry","category-llm","entry"],"jetpack_publicize_connections":[],"jetpack_featured_media_url":"","jetpack_shortlink":"https:\/\/wp.me\/p1q6tG-3n2","jetpack_sharing_enabled":true,"jetpack_likes_enabled":true,"_links":{"self":[{"href":"https:\/\/ihower.tw\/blog\/wp-json\/wp\/v2\/posts\/12960","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/ihower.tw\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/ihower.tw\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/ihower.tw\/blog\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/ihower.tw\/blog\/wp-json\/wp\/v2\/comments?post=12960"}],"version-history":[{"count":51,"href":"https:\/\/ihower.tw\/blog\/wp-json\/wp\/v2\/posts\/12960\/revisions"}],"predecessor-version":[{"id":13046,"href":"https:\/\/ihower.tw\/blog\/wp-json\/wp\/v2\/posts\/12960\/revisions\/13046"}],"wp:attachment":[{"href":"https:\/\/ihower.tw\/blog\/wp-json\/wp\/v2\/media?parent=12960"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/ihower.tw\/blog\/wp-json\/wp\/v2\/categories?post=12960"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/ihower.tw\/blog\/wp-json\/wp\/v2\/tags?post=12960"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}