Counterspeech

Suite of LLaMA-2 models fine-tuned for the personalized generation of counter-narratives in response to toxic comments posted on Reddit.