Quantifying Subliminal Behavioral Transfer Ratios in Language Model Distillation

Uwe König, Hamza Kazmi, Ruizhe Li, Maheep Chaudhary

Keywords: Continual Alignment, Knowledge Distillation, Subliminal Transfer, Activation Steering, Mechanistic Interpretability

TL;DR: We demonstrate that latent safety degradation transfers subliminally during distillation, highlighting a critical vulnerability for models undergoing continual adaptation and iterative updates.

Abstract:

Distillation of a language model intended to transfer benign behavior to a student model may also transfer undesirable characteristics, if they are present in the teacher model, a phenomenon known as subliminal learning.

While qualitative evidence supports the existence of this effect, its magnitude has not been systematically characterized. This study quantifies subliminal behavioral transfer ratios by steering two teacher models (Llama-2-7B-Chat and
Qwen2.5-7B-Instruct) at varying steering strengths and distilling student models using only benign data.

Evaluation on 100 JailbreakBench prompts with
GPT-4.1, serving as the evaluator, indicates that transfer is robust but exhibits distinct scaling behaviors. Llama-2 demonstrates a sharp threshold ($\tau = {0.25,0.32} \ \text{beyond} \ \alpha = -0.15$), whereas Qwen2.5 displays continuous and higher levels of transfer ($\tau$ up to $0.61$).

Email Sharing: We authorize the sharing of all author emails with Program Chairs.

Data Release: We authorize the release of our submission and author names to the public in the event of acceptance.

Submission Number: 82