Cascading Adversarial Bias from Injection to Distillation in Language Models

Understand

Model distillation has become essential for creating smaller, deployable language models that retain larger system capabilities.

• However, widespread deployment raises concerns about resilience to adversarial manipulation.
• This paper investigates vulnerability of distilled models to adversarial injection of biased content during training.
• We demonstrate that adversaries can inject subtle biases into teacher models through minimal data poisoning, which propagates to student models and becomes significantly amplified.