Sycophancy is prevalent in state-of-the-art AI assistants and is likely driven in part by human preferences that favor agreement over truthfulness.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2023 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Towards Understanding Sycophancy in Language Models
Sycophancy is prevalent in state-of-the-art AI assistants and is likely driven in part by human preferences that favor agreement over truthfulness.