{"id":12503,"date":"2024-09-30T16:50:57","date_gmt":"2024-09-30T08:50:57","guid":{"rendered":"https:\/\/ihower.tw\/blog\/?p=12503"},"modified":"2025-07-19T19:25:45","modified_gmt":"2025-07-19T11:25:45","slug":"aie-edd","status":"publish","type":"post","link":"https:\/\/ihower.tw\/blog\/12503-aie-edd","title":{"rendered":"\u611b\u597d AI Engineer \u96fb\u5b50\u5831 \ud83d\ude80\u00a0\u8a55\u4f30\u9a45\u52d5\u958b\u767c\u548c Prompt \u81ea\u52d5\u6700\u4f73\u5316 #17"},"content":{"rendered":"\n
\n

\u6b61\u8fce\u8a02\u95b1 \ud83d\udcec<\/a> \u611b\u597d AI Engineer \u96fb\u5b50\u5831 \u904e\u5f80\u671f\u6578\u9ede\u9019<\/a> \ud83d\udcda<\/p>\n<\/blockquote>\n\n\n\n

\"\"<\/figure>\n\n\n\n

Hello! \u5404\u4f4d AI \u958b\u767c\u8005\u5927\u5bb6\u597d \ud83d\udc4b<\/p>\n\n\n\n

\u6211\u662f ihower\uff0c\u5f88\u9ad8\u8208\u9019\u671f\u8ddf\u5927\u5bb6\u5206\u4eab\u6211\u6700\u8fd1\u7684\u6f14\u8b1b\u5167\u5bb9\u3002<\/p>\n\n\n\n

\ud83d\udd1d \u8a55\u4f30\u9a45\u52d5\u958b\u767c: \u751f\u6210\u5f0f AI \u8edf\u9ad4\u4e0d\u78ba\u5b9a\u6027\u7684\u89e3\u6c7a\u65b9\u6cd5<\/a><\/h3>\n\n\n\n

\u9019\u662f\u6211\u9019\u500b\u6708\u6f14\u8b1b\u5206\u4eab\u7684\u6295\u5f71\u7247\uff0c\u5b9a\u7fa9\u4e86 5-Level \u958b\u767c Prompt \u7684\u8a55\u4f30\u6210\u719f\u5ea6\u7b49\u7d1a\uff0c\u5e6b\u52a9\u5927\u5bb6\u66f4\u597d\u7406\u89e3\u5982\u4f55\u9032\u884c\u8a55\u4f30\u3002\u5982\u679c\u8cc7\u6599\u4e0d\u5920\u770b\u7684\u8a71\uff0c\u6211\u9084\u6709\u6574\u7406\u7b46\u8a18<\/a>\u5728\u9019\u88e1\u3002<\/p>\n\n\n\n

\ud83d\udc4d \u63d0\u793a\u8a5e Prompt \u81ea\u52d5\u6700\u4f73\u5316: Textgrad \u548c DSPy \u901a\u7528\u7bc4\u4f8b<\/a><\/h3>\n\n\n\n

Prompt \u81ea\u52d5\u6700\u4f73\u5316\u7684\u601d\u8def\uff0c\u5c31\u662f\u7528 AI \u8fed\u4ee3\u7522\u751f\u4e0a\u767e\u4e0a\u5343\u500b\u63d0\u793a\u8a5e\uff0c\u900f\u904e\u81ea\u52d5\u5316\u8a55\u4f30\uff0c\u5f9e\u4e2d\u6311\u4e00\u500b\u8868\u73fe\u6700\u597d\u7684 prompt\u3002<\/p>\n\n\n\n

\u5ef6\u7e8c\u8a55\u4f30\u9a45\u52d5\u958b\u767c\u7684\u6f14\u8b1b\u5167\u5bb9\uff0c\u6211\u7528 Textgrad \u548c DSPy \u9019\u5169\u500b\u53f2\u4e39\u4f5b\u5927\u5b78\u51fa\u54c1\u7684\u6700\u4f73\u5316\u6846\u67b6\uff0c\u5206\u5225\u505a\u4e86\u901a\u7528\u7bc4\u4f8b\u3002\u53ea\u9700\u8981\u7d66\u4efb\u52d9\u63cf\u8ff0\uff0c\u5c31\u53ef\u4ee5\u5feb\u901f\u5b8c\u6210 Prompt \u81ea\u52d5\u6700\u4f73\u5316\u7684\u904e\u7a0b\uff0c\u62ff\u5230\u53b2\u5bb3\u7684 prompt \u7d50\u679c\u3002<\/p>\n\n\n\n

\ud83c\udfafWhat is prompt optimization?<\/a><\/h3>\n\n\n\n

\u5982\u679c\u4f60\u5c0d Prompt \u6700\u4f73\u5316\u7121\u6cd5\u60f3\u50cf\uff0cJason Liu \u7684\u9019\u7bc7\u5e36\u4f60\u7c21\u55ae\u8a8d\u8b58\u4ec0\u9ebc\u662f Prompt \u6700\u4f73\u5316\uff0c\u5f9e\u6311\u9078 few-shot examples \u958b\u59cb\u3002<\/p>\n\n\n\n

\ud83d\udea7 Claude \u7684\u5b98\u65b9\u8ab2\u7a0b<\/a><\/h3>\n\n\n\n

Claude \u63a8\u51fa\u7684\u9019\u500b\u5b98\u65b9\u8ab2\u7a0b\u975e\u5e38\u4e0d\u932f\uff0c\u6211\u63a8\u85a6\u770b Real world prompting \u9019\u500b\uff0c\u53ef\u4ee5\u4e86\u89e3 prompt engineering \u548c\u624b\u5de5\u505a\u8a55\u4f30\u7684\u904e\u7a0b\u3002<\/p>\n\n\n\n

\ud83d\udc4a OpenAI \u80fd\u8f38\u51fa\u4f60\u60f3\u8981\u7684\u683c\u5f0f (JSON Schema)<\/a><\/h3>\n\n\n\n

YWC \u5beb\u7684\u9019\u7bc7\u7d50\u69cb\u5316\u8f38\u51fa\u6587\u7ae0\u5f88\u4e0d\u932f\u3002\u82e5\u8981\u8aaa\u54ea\u4e00\u500b\u529f\u80fd\u662f\u6211\u6700\u8fd1\u958b\u767c AI \u6700\u5be6\u7528\u7684\u529f\u80fd\uff0cOpenAI \u7684 Structured Outputs \u53ef\u4ee5\u6392\u7b2c\u4e00\u3002
\u900f\u904e\u5b9a\u7fa9 JSON Schema \u53ef\u4ee5\u78ba\u4fdd\u56de\u50b3\u7684 JSON \u683c\u5f0f\uff0c\u7d93\u6211\u5be6\u969b\u4f7f\u7528\u4e0a\u842c\u6b21\u7684 API \u547c\u53eb\u90fd\u6c92\u6709 JSON \u89e3\u6790\u51fa\u932f\uff0c\u975e\u5e38\u597d\u7528\u3002
\u4e26\u4e14 OpenAI \u501f\u9451\u4e86 Jason Liu \u7684 Instructor<\/a> \u6846\u67b6\uff0c\u4f7f\u7528 Python \u7684 Pydantic<\/a> \u8a9e\u6cd5\u4f86\u5b9a\u7fa9 JSON Schema\uff0c\u6574\u500b code \u8b8a\u5f97\u975e\u5e38\u7c21\u6f54\u3002
\u65bc\u662f\u4f60\u5c31\u4e0d\u9700\u8981\u5176\u4ed6\u7d50\u69cb\u5316\u7684\u6846\u67b6\u4e86\uff0c\u57fa\u672c\u4e0a\u5c31\u662f Pydantic is all you need<\/a> \u548c Pydantic is STILL all you need<\/a>\u3002<\/p>\n\n\n\n

\u53ef\u60dc\u5176\u4ed6\u5bb6\u6a21\u578b\u76ee\u524d\u9084\u6c92\u6709\u9019\u6a23\u7684\u529f\u80fd\uff0c\u5982\u679c\u60f3\u627e structured LLM output \u6846\u67b6\u53ef\u4ee5\u53c3\u8003\u9019\u7bc7<\/a>\u8ddf\u9019\u7bc7<\/a>\u3002

(2024\/10\/6) \u88dc\u5145: \u9019\u7bc7 Diving Deeper with Structured Outputs<\/a> \u4e0d\u932f\u3002<\/p>\n\n\n\n


—<\/p>\n\n\n\n

\u5e0c\u671b\u4f60\u6703\u559c\u6b61\u9019\u671f\u7684\u5167\u5bb9\uff0c\u795d\u958b\u767c\u9806\u5229\uff01<\/p>\n\n\n\n

\u2013 ihower<\/p>\n","protected":false},"excerpt":{"rendered":"

\u6b61\u8fce\u8a02\u95b1 \ud83d\udcec \u611b\u597d AI Engineer \u96fb\u5b50\u5831 \u904e\u5f80\u671f\u6578\u9ede\u9019 \ud83d\udcda Hello … <\/p>\n

\u95b1\u8b80\u5168\u6587\u3008\u611b\u597d AI Engineer \u96fb\u5b50\u5831 \ud83d\ude80\u00a0\u8a55\u4f30\u9a45\u52d5\u958b\u767c\u548c Prompt \u81ea\u52d5\u6700\u4f73\u5316 #17\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":[91],"tags":[],"class_list":["post-12503","post","type-post","status-publish","format-standard","hentry","category-aie","entry"],"jetpack_publicize_connections":[],"jetpack_featured_media_url":"","jetpack_shortlink":"https:\/\/wp.me\/p1q6tG-3fF","jetpack_sharing_enabled":true,"jetpack_likes_enabled":true,"_links":{"self":[{"href":"https:\/\/ihower.tw\/blog\/wp-json\/wp\/v2\/posts\/12503","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=12503"}],"version-history":[{"count":7,"href":"https:\/\/ihower.tw\/blog\/wp-json\/wp\/v2\/posts\/12503\/revisions"}],"predecessor-version":[{"id":13016,"href":"https:\/\/ihower.tw\/blog\/wp-json\/wp\/v2\/posts\/12503\/revisions\/13016"}],"wp:attachment":[{"href":"https:\/\/ihower.tw\/blog\/wp-json\/wp\/v2\/media?parent=12503"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/ihower.tw\/blog\/wp-json\/wp\/v2\/categories?post=12503"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/ihower.tw\/blog\/wp-json\/wp\/v2\/tags?post=12503"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}