-
may research note: mapping the space of LLM values
predicting alignment generalization and taxonomizing LLM values
-
when llms can write fiction, how will we know?
on evals of subjective tasks
predicting alignment generalization and taxonomizing LLM values
on evals of subjective tasks