argilla/ultrafeedback-binarized-preferences-cleaned-kto
Viewer
•
Updated
•
231k
•
88
•
6
This collection contains a list of curated preference datasets for KTO fine-tuning for intent alignment of LLMs through signals.
Note KTO transformed version of "argilla/ultrafeedback-binarized-preferences-cleaned".
Note KTO transformed version of "argilla/distilabel-intel-orca-dpo-pairs"
Note KTO transformed version of "argilla/distilabel-capybara-dpo-7k-binarized".
Note KTO transformed version of "argilla/dpo-mix-7k".