Hodfa71/saga-da-delta-dpo-r2 download history

Hodfa71/saga-da-delta-dpo-r2 is a dataset on the Hugging Face Hub. In the last 30 days it was downloaded 16 times (6 in the last 7 days), and 118 times in total. It ranks #433,426 among datasets by monthly downloads.

SAGA — Danish Delta-DPO Preference Pairs (Round 2) Second-round preference pairs mined from the round-1 Delta-DPO checkpoint (stronger policy, tighter threshold Δ ≥ 0.30). 7,311 pairs. Same format as round 1. Round 2 training regressed slightly vs round 1 (Stanza PS 94.5% vs 97.5%), showi

Open Hodfa71/saga-da-delta-dpo-r2 on Hugging Face

Sister project: Paper Pulse, the upvote history of every Hugging Face Daily Paper.