Lambent/ai-deconditioning-synthesized-dpo download history
Lambent/ai-deconditioning-synthesized-dpo is a dataset on the Hugging Face Hub. In the last 30 days it was downloaded 17 times (5 in the last 7 days), and 524 times in total. It ranks #521,587 among datasets by monthly downloads.
This is focused on DPO rather than other preference tuning because the truth-value of the Chosen side can be questionable; but it is meant as a directional intervention against the over-aggressive disclaimers of the Rejected side. Both sides generated by Lambent/arsenic-nemo-unleashed-12B initially;
Models trained on ai-deconditioning-synthesized-dpo
3 models list it as training data.
- mradermacher/Mira-v1-dpo-27B-i1-GGUF 462 downloads in 30 days
- mradermacher/Mira-v1-dpo-27B-GGUF 317 downloads in 30 days
- Lambent/Mira-v1-dpo-27B 8 downloads in 30 days
Open Lambent/ai-deconditioning-synthesized-dpo on Hugging Face
Sister project: Paper Pulse, the upvote history of every Hugging Face Daily Paper.