Datasets:
OALL
/

Modalities:
Text
Formats:
parquet
ArXiv:
Libraries:
Datasets
pandas

A bug in the first row of test split in `mcq_exams_test_ar` subset

#2
by sadra-barikbin - opened

Hi there!
There's seemingly a bug in the first row of the test split in mcq_exams_test_ar subset. The second choice is the correct one while the label is 1.

sadra-barikbin changed discussion title from a bug in the first row of test split in `mcq_exams_test_ar` subset to A bug in the first row of test split in `mcq_exams_test_ar` subset
Open Arabic LLM Leaderboard org

Dear @sadra-barikbin , Thank you do much for taking the time to report this, we appreciate that a lot πŸ€—.
As for your question, the indexing start with 0, So as the second choice is the correct one, then the index 1 is correct. For a better idea about the overall structure and indexes please check the test section.
Thank you again, and please take the time to spot any errors that might be in one of the benchmarks and let us know πŸ€—

Ali-C137 changed discussion status to closed

Sign up or log in to comment