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.

Long-form factuality in large language models

18 pointsby rootforceabout 1 year ago

2 comments

rstabout 1 year ago
Hmmm... checking against external sources is an interesting idea -- but using Google as a source of ground truth is a little bit tricky, given how often these days Google itself is spitting up confabulated AI-generated crud (or other low-quality stuff).
评论 #39873244 未加载
cl42about 1 year ago
For those interested in using search-augmented &quot;reasoning&quot;, I implemented something similar in Emerging Trajectories[1], an open source package that forecasts geopolitical and economic events. We extract facts[2] from various websites (Google searches, news articles, RSS feeds) and have the LLM generate a hypothesis on a metric.<p>We&#x27;re tracking the info forecasts to see how well this does for future events. For example, we&#x27;re pitting the LLMs against each other to predict March 2024 CPI[3].<p>[1] <a href="https:&#x2F;&#x2F;emergingtrajectories.com&#x2F;" rel="nofollow">https:&#x2F;&#x2F;emergingtrajectories.com&#x2F;</a><p>[2] Sample code: <a href="https:&#x2F;&#x2F;github.com&#x2F;wgryc&#x2F;emerging-trajectories&#x2F;blob&#x2F;main&#x2F;emergingtrajectories&#x2F;factsforecaster.py">https:&#x2F;&#x2F;github.com&#x2F;wgryc&#x2F;emerging-trajectories&#x2F;blob&#x2F;main&#x2F;eme...</a><p>[3] <a href="https:&#x2F;&#x2F;emergingtrajectories.com&#x2F;a&#x2F;statement&#x2F;28" rel="nofollow">https:&#x2F;&#x2F;emergingtrajectories.com&#x2F;a&#x2F;statement&#x2F;28</a>