fascinating paper — it explores how LLMs take different sides on various geop...
View original threadfascinating paper — it explores how LLMs take different sides on various geopolitical disputes based on what language they’re speaking in
@jmsdnns.bsky.social
@timkellogg.me I can't remember where, but I recall you recently discussing models reasoning in non-english languages.
I came across a paper that suggests models have different biases depending on what language they're using. Interesting implications!
www.cis.upenn.edu/~ccb/publica...
I came across a paper that suggests models have different biases depending on what language they're using. Interesting implications!
www.cis.upenn.edu/~ccb/publica...
40
8
imo this is an artifact of LLMs goal of “talking like a person would”
e.g. an LLM isn’t “left leaning”, it leans in all directions, you just observed it mimicking a person who is left-leaning
in this case, changing the language elicits different target personalities to mimic
e.g. an LLM isn’t “left leaning”, it leans in all directions, you just observed it mimicking a person who is left-leaning
in this case, changing the language elicits different target personalities to mimic
9