When Backdoors Speak: Understanding LLM Backdoor Attacks Through Model-Generated Explanations

Open in new window