Skip to content

Large Language Models Reflect the Ideology of Their Creators

arxiv.org
4 pointsdelian662 comments
On HN

Comments

This article conflates China and the US as countries in which LLMs are built with the corpus or body of text that was used in training.

The creator of an LLM is the team that writes the GPT code, trains it on a body of text, and creates a UI to allow prediction or other output. The "ideology" in those texts belongs somewhat to the people who selected them but mostly to the people who wrote them.

It doesn't take science to understand that.

LLMS are a reflection of us.

“Mirror, Mirror on the wall, who's the fairest of them all?”

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.