goal : conversation (english & french)
uncensored : lightly
base : Nemotron and Llama 3.3
Trained on exclusively english dataset, while trying to keep all the french talk.
DPO made on human data vs llm output
gated for now, just apply, I will accept. testing shiet