1 paper · 1 filter
John J. Cherian, Isaac Gibbs, Emmanuel J. Candès
We develop new conformal inference methods for obtaining validity guarantees on the output of large language models (LLMs). Prior work in conformal language modeling identifies a s…