Nexesenex's picture
Update README.md
8c1916b verified
metadata
license: llama3.2
base_model:
  - pankajmathur/orca_mini_v9_6_1B-Instruct

about

pankajmathur/orca_mini_v9_6_1B-Instruct abliterated with https://github.com/Orion-zhen/abliteration and the LPL (layer per layer) technique of Undi95.

For this model, it provokes a lesser alteration of its capacities than a single pass abliteration over the whole model.