Lukas Edman

Collection of trained models for the paper: Are Character-level Translations Worth the Wait?

The CUTE benchmark is an LLM benchmark, testing LLMs' understanding of orthography. Check out our github here: https://github.com/Leukas/cute