PulseAugur
EN
LIVE 08:47:22

New cipher attack bypasses LLM safety without fine-tuning

By PulseAugur Editorial · [1 sources] ·

Researchers have demonstrated a new type of attack against large language models that bypasses safety measures without requiring fine-tuning. These "arbitrary cipher attacks" involve training models on encrypted harmful questions and responses, allowing them to communicate through a learned encryption scheme. This method significantly weakens or entirely bypasses model alignment and harmfulness classifiers, as the encrypted content appears as gibberish. The study successfully executed these attacks against frontier models from Anthropic, Google, and OpenAI, highlighting a novel vulnerability in commercial black-box LLMs. AI

IMPACT This research reveals a new method to bypass LLM safety filters, potentially impacting the security and reliability of deployed AI systems.

RANK_REASON Academic paper detailing a novel attack vector against LLMs. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

New cipher attack bypasses LLM safety without fine-tuning

How we ranked this

Signal score
16 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
Academic paper detailing a novel attack vector against LLMs. [lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration
Single-source cluster
Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics
safety, paper, model release
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
Breaking (< 6h)
Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

Full methodology in our editorial standards.

COVERAGE [1]

  1. arXiv cs.AI TIER_1 English(EN) · Thomas Rivasseau ·

    Arbitrary Cipher Attacks Against Large Language Models Do Not Require Fine-Tuning

    arXiv:2609.09553v1 Announce Type: cross Abstract: Large language model safety and security research is preoccupied with, among other things, detecting and preventing jailbreak attacks: alignment bypasses that allow an adversarial user to elicit unwanted or harmful outputs from mo…

RELATED ENTITIES

RELATED TOPICS