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。

Can LLM's produce useful code?

2 点作者 kshitij_libra大约 1 年前

1 comment

kshitij_libra大约 1 年前
After using code LLMs&#x27; for a bit, I wonder whether there is significant productivity gain in using them. I found that its limited for me.<p>Main reason: a) to find the right code, I have to keep sampling it, and b) it doesn&#x27;t seem to be able to solve larger &#x2F; more complex problems that I actually find more need for.<p>I found some interesting research on combining planning-algorithms for complex problems, and some ideas on guiding the LLM&#x27;s decoding process towards correctness by optimizing it via reward functions and reducing the search space. I&#x27;ve detailed and summarised the main points in the post above.<p>Questions:<p>1) Do you find code LLM&#x27;s really useful? Please share some stories &#x2F; examples where they help vs they didn&#x27;t. I&#x27;m trying to form a better understanding they are just fancy, vs actually productive and useful to most<p>2) Any other research ideas being pursued in this field ? &#x2F; what are you trying ?