1 paper · 1 filter
Tanush Chopra, Michael Li, Jacob Haimes
When large language models (LLMs) are asked to perform certain tasks, how can we be sure that their learned representations align with reality? We propose a domain-agnostic framewo…