TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

Lumos: Learning Agents with Unified Data, Modular Design, and Open-Source LLMs

102 pointsby milliondreamsabout 1 year ago

6 comments

CGamesPlayabout 1 year ago
Random, off-topic observation: it seems like I see lots of open-source LLM prompts have prompts that aren&#x27;t proofread. For example [0], &quot;ounch&quot;, &quot;he spend&quot;, general lack of punctuation, and so on. But the big model providers don&#x27;t do this [1], so I&#x27;m curious how much impact proper grammar in prompts makes (and if it&#x27;s a positive or negative impact).<p>[0] <a href="https:&#x2F;&#x2F;github.com&#x2F;allenai&#x2F;lumos&#x2F;blob&#x2F;main&#x2F;data&#x2F;incontext.py#L91">https:&#x2F;&#x2F;github.com&#x2F;allenai&#x2F;lumos&#x2F;blob&#x2F;main&#x2F;data&#x2F;incontext.py...</a><p>[1] <a href="https:&#x2F;&#x2F;twitter.com&#x2F;AmandaAskell&#x2F;status&#x2F;1765207842993434880" rel="nofollow">https:&#x2F;&#x2F;twitter.com&#x2F;AmandaAskell&#x2F;status&#x2F;1765207842993434880</a>
评论 #39904093 未加载
评论 #39903879 未加载
antupisabout 1 year ago
I think that to get truly working agents we need huge working context windows like 1M - 100M tokens everything before that will be kinda hackish.
评论 #39903697 未加载
评论 #39902936 未加载
zzzzzzzzzz10about 1 year ago
How can I run this locally does anyone know? Can&#x27;t find any instructions.
评论 #39906546 未加载
dvtabout 1 year ago
The paper[1] seems promising, is there a fine-tuned model available? Or will we have to fine tune Llama-7b or Mistral-7b ourselves?<p>[1] <a href="https:&#x2F;&#x2F;arxiv.org&#x2F;pdf&#x2F;2311.05657.pdf" rel="nofollow">https:&#x2F;&#x2F;arxiv.org&#x2F;pdf&#x2F;2311.05657.pdf</a>
jondwillisabout 1 year ago
All of these papers and projects need less jargon, and more examples and proof of performance.
milliondreamsabout 1 year ago
Looks promising approach to Agentic AI systems.