Home » Lakera launches to protect large language models from malicious prompts