Datasets:

Modalities:
Text
Formats:
parquet
Languages:
Spanish
Size:
< 1K
ArXiv:
License:

Update ARC ES validation splits in Parquet

#1
Language Technologies Laboratory @ Barcelona Supercomputing Center org
edited Jun 1

Update the repo structure to make the dataset cleaner in the viewer

Language Technologies Laboratory @ Barcelona Supercomputing Center org

The previous version was stored as JSONL files, so ARC-Challenge and ARC-Easy were not properly reflected as separate validation splits in the Dataset Viewer. This update converts both files to Parquet using the standard ARC format and separates them into their own validation configs

gonzalez-agirre changed pull request status to merged
gonzalez-agirre deleted the refs/pr/1 ref

Sign up or log in to comment