CoolFace
Datasetpublic

ceselder/loracle-pretrain

loracle-pretrain-qa-v4.1-25k v4.1 pretraining dataset for the LoRACLE. Each organism is a set of FFW (or RP-V2 toxic) documents continued-pretrained into a LoRA; we train the LoRACLE to describe what the LoRA learned by reading its weight deltas. Exactly 2 rows per organism (Slot A + Slot B). ~5% toxic organisms (2.5% sporadic + 2.5% all_toxic) for robustness training.

sourceHugging Faceupdated 5mo agoView on Hugging Face
0likes6downloads

ceselder/loracle-pretrain · main · files are served by the source, never re-hosted here