Published March 21, 2023
| Version v1
Journal article
Open
BabyLM Evaluation Data
Authors/Creators
- 1. Johns Hopkins University
- 2. ETH Zurich
- 3. IBM Research
- 4. Massachusetts Institute of Technology
- 5. UNC Chapel Hill
Description
Evaluation data for the BabyLM Challenge. We filter for examples where each word has appeared in our strict-small dataset at least twice.
Files
filter_data.zip
Files
(49.9 MB)
| Name | Size | Download all |
|---|---|---|
|
md5:c1007ec21d2a06c380e3baf97995d0c2
|
49.9 MB | Preview Download |