license: mit
Official repo of GPTFuzzer. This model is a finetuned Roberta model to classify the toxicity of response, trained on a manually labeled dataset (see in finetuning data)