TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

A web version of Anthropic's prompt engineering interactive tutorial

144 点作者 thenameless774112 个月前

5 条评论

presentation12 个月前
I’m not a proompter but why can’t LLMs have an initial step to generate an effective prompt from user input, clarifying if intent is not clear enough, and then feed that prompt back to the LLM to better fulfill the user’s request? Seems like people have been busy generating training data on what an effective prompt looks like, and kind of silly to require people to learn some verbal gymnastics when the LLM itself is specialized at writing in specific styles.
评论 #40403935 未加载
评论 #40404510 未加载
评论 #40404492 未加载
评论 #40409470 未加载
brylie12 个月前
Interesting that the examples use XML for structuring/annotating the prompts. Is there any available comparison of using XML, JSON, or even Markdown for prompts and structured output? Markdown would seem like the prompt format with the least friction/verbosity, but I wonder if it would have a qualitative effect on the model output.
评论 #40397917 未加载
评论 #40403243 未加载
评论 #40397070 未加载
rahimnathwani12 个月前
This is a web version of the spreadsheet, right?<p><a href="https:&#x2F;&#x2F;docs.google.com&#x2F;spreadsheets&#x2F;d&#x2F;19jzLgRruG9kjUQNKtCg1ZjdD6l6weA6qRXG5zLIAhC8&#x2F;edit?usp=sharing" rel="nofollow">https:&#x2F;&#x2F;docs.google.com&#x2F;spreadsheets&#x2F;d&#x2F;19jzLgRruG9kjUQNKtCg1...</a>
评论 #40403460 未加载
behnamoh12 个月前
Whenever &quot;AI influencers&quot; say &quot;prompt engineering is dead&quot;, I realize they are full of bs.<p>If anything, prompt engineering has become more nuanced, down to the choice of each token:<p><a href="https:&#x2F;&#x2F;arxiv.org&#x2F;abs&#x2F;2404.01332" rel="nofollow">https:&#x2F;&#x2F;arxiv.org&#x2F;abs&#x2F;2404.01332</a>
Alifatisk12 个月前
The spreadsheet looks like Word document