Lisa is a series of models fine-tuned using the dataset Lisa.csv.

Lisa-8b is derived from "Goekdeniz-Guelmez/Josiefied-Qwen3-8B-abliterated-v1".

This is an experimental early test version.

Training scripts / settings / dataset / explanation / etc. here.

Reasoning is scuffed. Perhaps a feature?

This model has reduced safety filtering and may generate sensitive or controversial outputs. Use responsibly and at your own risk.