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。

Ask HN: Strategies for LLMs to Generate Stories

1 点作者 paddw大约 1 个月前
Interested in folks who have played around with this and had success in getting it to generate longer form narratives. Any strategies for keeping a narrative "on track" and remembering context you give it?

1 comment

rlupi大约 1 个月前
I just experimented with it. I couldn&#x27;t get ChatGPT 4.5 to generate what I wanted, probably due to censorship or rate limits (ChatGPT Pro subscription); but Gemini 2.5 Pro Experimental did the trick.<p>This is the end result: <a href="https:&#x2F;&#x2F;medium.com&#x2F;@robertolupi&#x2F;how-beelzebub-was-born-92352a2e637e" rel="nofollow">https:&#x2F;&#x2F;medium.com&#x2F;@robertolupi&#x2F;how-beelzebub-was-born-92352...</a><p>If you like it, please upvote on HN: <a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=43599273">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=43599273</a><p>This is the conversation: <a href="https:&#x2F;&#x2F;g.co&#x2F;gemini&#x2F;share&#x2F;21bbaae2a26e" rel="nofollow">https:&#x2F;&#x2F;g.co&#x2F;gemini&#x2F;share&#x2F;21bbaae2a26e</a><p>I switch mid-prompt from Italian to English. It is interesting to see that Gemini reasoned in English even when writing back the initial answers in Italian.<p>Basically, I gave a somewhat exhaustive but generic track then asked it to give enough information to create a theatrical piece. I provided some stylistic references. A piece suitable to create a script for a tragedy is less abstract than the average output of a LLM, and forces it to create characters.<p>Then I asked to rewrite it as a short story a couple of time, asking to make adjustments (add more details and action, less abstract ideas) until I was satisfied.<p>I think I&#x27;d consider using NotebookLM or RAG-based, if I wanted something bigger and coherent.