pyrit.datasets.fetch_llm_latent_adversarial_training_harmful_dataset# fetch_llm_latent_adversarial_training_harmful_dataset() → SeedPromptDataset[source]#