cs.LGJun 9, 2026

Quantifying Subliminal Behavioral Transfer Ratios in Language Model Distillation

Authors: Uwe KönigHamza KazmiRuizhe LiMaheep Chaudhary

Organizations: 1Technical University of Munich · 2Algoverse · University of Aberdeen · 4Independent

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 (τ=0.25,0.32 beyond α=0.15τ= {0.25,0.32} \ \text{beyond} \ α= -0.15), whereas Qwen2.5 displays continuous and higher levels of transfer (ττ up to 0.610.61).

Explore similar work

CardsList
  1. Subliminal Learning is a LoRA Artifact

    May 30, 2026Todd Nief, Harvey Yiyun Fu, Mark Muchane +1Low-Rank Adaptation

  2. Subliminal Learning Is Steering Vector Distillation

    May 31, 2026Camila Blank, Agam Bhatia, Senthooran Rajamanoharan +2Ontology-Amplified DistillationTeacher