More human than human: measuring ChatGPT political bias #ai #Science #politics https://t.co/X0hLw1zRuf
So far, Altmetric has seen
207 X posts from 194 X users, with an upper bound of 1,041,889 followers.
38 followers
1,055 followers
The ideological biases in LLM are well-documented and represent the general values of the firm that fine-tuned the base model. Meta's LLAMA model has a right-wing authoritarian bias, and OpenAI's GPT family, including ChatGPT, has a left-libertarian bias.
1,006 followers
RT @jmhorp: Here's a new paper in @_PublicChoice on ChatGPT and political bias: "We find robust evidence that ChatGPT presents a significa…
198 followers
RT @jmhorp: Here's a new paper in @_PublicChoice on ChatGPT and political bias: "We find robust evidence that ChatGPT presents a significa…