Lukas Edman
leukas
AI & ML interests
NLP, tokenization, pretraining, low-resource
Recent Activity
new activity 13 days ago
leukas/climbmix-4b-shuffle:[bot] Conversion to Parquet new activity 13 days ago
leukas/climbmix-1b-shuffle:[bot] Conversion to Parquet updated a dataset 14 days ago
leukas/climbmix-1b-shuffleOrganizations
BabyLM2024 models
The CUTE benchmark is an LLM benchmark, testing LLMs' understanding of orthography. Check out our github here: https://github.com/Leukas/cute