---
title: "Discussion on \"A Multidisciplinary Theory of Exploiting LLMs  and Guardrails: The Collapse of Embedding Geometry\" | Hashnode"
description: "Discussion on \"A Multidisciplinary Theory of Exploiting LLMs  and Guardrails: The Collapse of Embedding Geometry\". Multidisciplinary red team research on LLM SecOps reveals new vector attacks: Adversarial Vector Semiotics, Epistemologic Poisoning, CryptoLinguistics"
url: "https://hashnode.com/posts/a-multidisciplinary-theory-of-exploiting-llms-and-guardrails-the-collapse-of-embedding-geometry/68d027855a74ece3f115183b"
---

[Feed](https://hashnode.com/feed) /

Discussion

# [A Multidisciplinary Theory of Exploiting LLMs and Guardrails: The Collapse of Embedding Geometry](https://ai-blog.dev/a-multidisciplinary-theory-of-exploiting-llms-and-guardrails-the-collapse-of-embedding-geometry?utm_source=hashnode&utm_medium=feed)

Multidisciplinary red team research on LLM SecOps reveals new vector attacks: Adversarial Vector Semiotics, Epistemologic Poisoning, CryptoLinguistics

Sep 21, 2025 · 13 min read · ai-blog.dev  
Read the article

-   [#ai](https://hashnode.com/tag/ai)
-   [#cybersecurity](https://hashnode.com/tag/cybersecurity)
-   [#aicybersecurity](https://hashnode.com/tag/aicybersecurity)
-   [#linguistics](https://hashnode.com/tag/linguistics)
-   [#guardrails](https://hashnode.com/tag/guardrails)

## Written by

[Amit Sides](https://hashnode.com/@amitsides)

@amitsides · Tel Aviv, Israel

  
[View profile](https://hashnode.com/@amitsides)

AI Engineer

Amit Sides is a AI Engineer, Python Backend Developer, DevOps Expert, DevSecOps &amp; MLOPS GITHUB https://github.com/amitsides

## Responses

Markdown supported · ⌘↵ to post

No responses yet.
