1 paper · 1 filter
Sebastian Pohl, Harsh Mehta, Pranav Mambayil +4
LLMs are increasingly deployed as proxies for human study participants in social science experiments, yet the fidelity of this practice has rarely been tested directly. We test whe…