Arabic ORPO-DPO Datasets - a arcee-globe Collection

arcee-globe 's Collections

Arabic ORPO-DPO Datasets

Multilingual LLM Papers

Arabic Light Benchmarks

Multilingual Datasets

Datasets with Educational Values

Arabic ORPO-DPO Datasets

updated 18 days ago

arcee-globe/cleaned-NoRobots-Command.R-ORPO

Viewer • Updated Aug 15, 2024 • 9.5k • 14

Note External Dataset
arcee-globe/filtered-arabic-distilabel-math-preference-orpo

Viewer • Updated Aug 15, 2024 • 1.96k • 6

Note Translated with GPT4o-mini
arcee-globe/mixed-argilla-orpo-mix-7k-arabic

Viewer • Updated Aug 15, 2024 • 6.75k • 7

Note External Dataset with mixture of multi-turn and single-turn samples
arcee-globe/arabic-jondurbin-truthy-orpo-v0.1

Viewer • Updated Aug 16, 2024 • 920 • 5

Note Translated with GPT4o-mini
arcee-globe/arabic-distilabel-capybara-orpo-7k-binarized

Viewer • Updated Aug 17, 2024 • 5.4k • 10

Note Has multi-turn instances Translated with GPT4o-mini
arcee-globe/arabic-orpo-dpo-mix-40k-filtered

Viewer • Updated Aug 17, 2024 • 31.8k • 4

Note Might have multi-turn instances Removed the toxic subset from the original dataset. Translated with GPT4o-mini