Myra Cheng

@myra.bsky.social

2908 Followers 147 Following

PhD candidate @ Stanford NLP https://myracheng.github.io/

Rightness judgment is higher and repair likelihood is lower for sycophantic AI
Description of Study 2 (hypothetical vignettes) and Study 3 (live interaction) where self-attributed wrongness and desire to initiate repair decrease, while response quality and trust increases.
Description of Study 1, where we characterize the prevalence of social sycophancy and find it to be highly prevalent across leading AI models
Screenshot of paper title: Sycophantic AI Decreases Prosocial Intentions and Promotes Dependence
Plots showing that DumT reduces MeanHumT and has higher performance on RewardBench than the baseline models.
human-like LLM outputs are strongly positively correlated with social closeness, femininity, and warmth (r = 0.87, 0.47, 0.45), and strongly negatively correlated with status (r = 0.80).
bar plot showing that human-likeness is lower in preferred responses
Screenshot of first page of the paper HumT DumT: Measuring and controlling human-like language in LLMs