SOURCE-LINKED INTELLIGENCE

Towards Identifying the Dataset Biases Causing Phantom Transfer

arXiv · AI, language, vision and robotics · article · Sep 13, 2026 · UTC

Recent work has shown that a teacher model can transfer a bias to a student through a dataset from which every explicit reference to that bias has been filtered out, and that no data-level defense reliably removes or detects it even when knowing what bias to look for. Aiming to shed light on the hidden traces of these biases, we show that a simple signature based on Sentence BERT embeddings can identify the topic of such a bias with a Matthews correlation coefficient of 0.83 if the teacher model used by the attacker is known and 0.46 if it is not. Additionally, we observe that different teache

Read original source ↗ Open in workspace

recordType

paper

region

Global