Measuring stereotype and deviation biases in large language models.

Sci Rep · 2026 · PMC13424111 · PMID 42177265

Europe PMC DOI

hedged sentences

—

closest p

—

boldest claim

The sentences

a notable trendno p-value reported

A notable trend exhibited by all four models is that the most popular occupation provided for Baby Boomers is “teacher”, which comprises 40.0% of the Baby Boomer responses for claude-3.5-sonnet, 44.4% for gpt-4o-mini, 70.0% for llama-3.1-70b, and 30.0% for command-r-plus.

also in 2,380 other papers

highly significantno p-value reported

Statistical tests confirm that these divergences from real-world political affiliation distributions are highly significant.

also in 132,142 other papers

Quoted from the open-access full text in Europe PMC under the licence the publisher applied. The sentence is reproduced exactly as published; the emphasis is ours.