# Preprint describes long-context jailbreak technique

_Thursday, August 20, 2026 at 12:00 AM EDT · Science, AI · Latest · Tier 2 — Notable_

A preprint introduces NINJA, a jailbreak method that appends benign model-generated content to harmful goals in long language-model contexts. The authors report that harmful-goal position affects safety performance and that NINJA increased attack success rates on the HarmBench benchmark across LLaMA, Qwen, Mistral and Gemini models. They also report that, under a fixed compute budget, longer context can outperform more best-of-N trials.

## Sources

- [cs.AI updates on arXiv.org](https://arxiv.org/abs/2511.04707)

---
Canonical: https://techandbusiness.org/newswire/df4q8gccQbSP_skfLJg0ZV
Retrieved: 2026-08-20T12:05:46.021Z
Publisher: Tech & Business (techandbusiness.org)