Publication Radar · Source metadata

Homeopathic Learning: Diluting poisoned fine-tuning data can strengthen its effect

Matt Clarke, Jan Dubi\'nski, James Chua, Simon Schrodi, Owain Evans, Alex Cloud

Publication record

Bibliographic evidence describes where and when a record appeared. It does not validate scientific claims.

First announced by arXiv: 2026-11-10 · Indexed by Pub-AI: 2026-11-10 UTC

preprint

Homeopathic Learning: Diluting poisoned fine-tuning data can strengthen its effect

Matt Clarke, Jan Dubi\'nski, James Chua, Simon Schrodi, Owain Evans, Alex Cloud

First announced by arXiv: 2026-11-10 · Original record ↗

arxiv
2611.10110
Explicit source versions (1)
  • v1 · Date not supplied
  1. Preprint available

    Source date: 2026-11-10 · First detected: 2026-11-10 UTC · Historical baseline

    Source evidence (1)
    • Official evidence ↗ · RSS guid=oai:arXiv.org:2611.10110v1; arxiv:announce_type=new; item pubDate (official metadata)
Abstract · arxiv

We study homeopathic learning, a phenomenon where poisoned training data with little effect on its own can create a hidden backdoor once clean, contrasting data is added. In our experiments, a model fine-tuned only on a "poison" dataset associated with a trait T (such as speaking German or being misaligned) does not learn T. Surprisingly, adding similarly formatted "clean" data not associated with T causes the model to learn T. This effect requires that the poison data have an identifying "trigger" feature that is absent from the clean data. We demonstrate this effect in a variety of settings: subliminal learning of backdoor behaviors, inoculation prompting of reward hacking transcripts, generalization from archaic bird names, and an MNIST classifier with mislabeled data. We propose a simple regression model as an explanation. In the poison data, the task and trigger always co-occur, and regularization favors attributing T to the task. Clean data without the trigger causes T to be attributed to the trigger instead. Consequently, the same training examples can have little effect as part of one training set but create a misalignment backdoor in another. This finding exposes a potential limitation of safety methods that audit or filter training data in isolation.