Morning Edition · Thursday, September 3, 2026Published at 2:26 AM EDT · New York

Researchers Recover Secrets From a Model's Hidden Context Without Any Jailbreak

The attack sends ordinary, non-malicious queries and uses a surrogate model to work out which candidate secret best explains the answers, which makes it hard to detect by prompt filtering.

A paper posted to arXiv, Context Inference Attacks Without Jailbreaks, describes a way to extract information from the hidden context an agentic system assembles before answering, without using adversarial prompts at all. Agentic deployment…