Etienne Casanova

On the Limits of LLM Adaptability: Impact of LLM Pre-Training on Annotation Task Performance

Pre-trained Large Language Models (LLMs) are increasingly used for zero-shot annotation and LLM-as-a-judge tasks, yet their reliability hinges on how pre-trained priors interact with user-provided instructions. We investigate three dimensions of this interaction: (1) how an LLM’s familiarity with da…