Merged Arabic clean dataset (fixed JSONL)This dataset merges:
Jianshu001/daily-conversation-batch-01-v5-clean first
Jianshu001/arabic-daily-v6-batch1to5-clean-nometa second
Notes:
Rewritten to strict JSONL to avoid dataset viewer parse errors.
Duplicate IDs skipped in the original merge build.
metadata removed.
