AI ethics evidence index Search
Source-linked daily record
Evidence record 13494 · automatically gathered

Geometric Configurations of Perturbed Jailbreak Prompts

Perturbation techniques that turn unsuccessful jailbreak prompts into successful ones are continuously evolving, constituting a major security threat to LLM safety. In this paper, we investigate the internal representations of such string-level perturbed jailbreak inputs in the small weight models of the Qwen-2.5-1.5B/-3B/-7B-Instruct and Llama-3.2-1B/-3B/-3.1-8B-Instruct families. We select two representation spaces: the last-layer-last-token embedding space and the top-50 next-token probabilit

Record details

Published: 22 July 2026
Source: arXiv red teaming query
Category: Research
Topics: Safety & alignment · Finance, VC & PE
Retrieved: 25 July 2026

source-onlyevidence status

These records share source-supplied organisations, an exact publisher byline, automatic topics or regions. The reason is shown on every link; related does not mean supporting, agreeing with or verifying this record.

Shared topic: Safety & alignment · Shared topic: Finance, VC & PE

Amplifying the storm: Climate disinformation dynamics during natural disasters on right-wing extremist Telegram channels

HKS Misinformation Review · 23 July 2026

Shared topic: Safety & alignment · Shared topic: Finance, VC & PE

Norm or Direction? Decoding Vision Mambas for High-Resolution Vision

arXiv · 21 July 2026

Shared topic: Safety & alignment · Shared topic: Finance, VC & PE

For What Reason? Interpreting Models' Encoding of Causation and Antithesis

arXiv · 20 July 2026

Shared topic: Safety & alignment · Shared topic: Finance, VC & PE

Autoregressive EHR Foundation Models with Multimodal Inputs

arXiv cs.LG · 24 July 2026

Shared topic: Safety & alignment · Shared topic: Finance, VC & PE

Internal Pluralism and the Limits of Pairwise Comparisons

arXiv cs.CY · 20 July 2026

Shared topic: Safety & alignment · Shared topic: Finance, VC & PE

An Iterative Geometric Approach to Optimizing Separating Hyperplanes

arXiv cs.LG · 19 July 2026

How to cite this record

ethics.ai (22 July 2026), “Geometric Configurations of Perturbed Jailbreak Prompts,” evidence record 13494, https://ethics.ai/record/13494 (originally published by arXiv red teaming query).

JSON

Use and limitations

This page is a stable index and citation surface for a source record. ethics.ai did not author the underlying report and has not independently verified every claim. Automatic topics may be imperfect. For consequential use, quote and cite the original publisher.