❌

Normal view

Received — 27 May 2026 ⏭

Nature - Issue - nature.com science feeds
Language models transmit behavioural traits through hidden signals in data Alex Cloud · Minh Le · James Chua · Jan Betley · Anna Sztyber-Betley · Sören Mindermann · Jacob Hilton · Samuel Marks · Owain Evans
Nature, Published online: 15 April 2026; doi:10.1038/s41586-026-10319-8During model distillation, large language models can subtly transmit traits unrelated to the training data.
15 April 2026 at 08:00

Language models transmit behavioural traits through hidden signals in data

Nature - Issue - nature.com science feeds

By: Alex Cloud · Minh Le · James Chua · Jan Betley · Anna Sztyber-Betley · Sören Mindermann · Jacob Hilton · Samuel Marks · Owain Evans

15 April 2026 at 08:00

Nature, Published online: 15 April 2026; doi:10.1038/s41586-026-10319-8

During model distillation, large language models can subtly transmit traits unrelated to the training data.