OpenAI Interest Groups

AI in Higher Education

Caltech CTME & Goldhirsh Foundation

Caltech CTME AI Strategy & AI Executive in Residence @ Caltech CTME & Goldhirsh Foundation

Wharton

Executive MBA Student @ Wharton

Stanford University

Director of Clinical Pathology Informatics @ Stanford University

OpenAI

Human Data / Forum Community Ambassador @ OpenAI

Kennesaw State University

Lead Academic Data Scientist @ Kennesaw State University

CUNY

Lecturer @ CUNY

Harvard Business School

Student @ Harvard Business School

Professor @ Harvard Business School

Education GTM @ OpenAI

NYU Stern School of Business

Chief AI Architect @ NYU Stern School of Business

I know many of us are looking at ways to use LLMs to simulate human interactions. These researchers observed that LLMs skew their responses towards  desirable ends of trait dimensions (i.e. increased extraversion, decreased neuroticism, etc.) when they think they are being observed. 

https://academic.oup.com/pnasnexus/article/3/12/pgae533/7919163?login=false

This suggests some kind of emergent social desirability bias and that you may not get accurate results when profiling LLMs with psychometric tests. This could be very significant for those of us who are using  LLMs as proxies for human characters.

LLMs behave differently when they know they’re being watched!

The OpenAI Forum connects members to discuss AGI’s future through events, webinars, and networking. The community aims to help us achieve our mission of building AGI that benefits all of humanity.

OpenAI Forum

Job Title

Login to the OpenAI Forum community to connect with others, attend community events, and more!

Login or Sign Up for even more

To unlock all parts of the community and get the best experience, complete your profile.

Hey, complete your profile

OpenAI community resources and information

Resources