Skip to main content
LLM Security Database
Skip to research search
Last analyzed 9/9/2026

Language Model Security Database

985 research findings · 1123 evaluated models

Filtered research findings

1 entry

Matches every word across titles, descriptions, sources, affected systems, and models.

Nonsensical CoT Reasoning

Published 5/1/2025
Analyzed 12/30/2025

Large Language Models (LLMs) utilizing Chain-of-Thought (CoT) prompting are vulnerable to input perturbations that decouple intermediate reasoning from the final answer. An attacker can generate adversarial examples using gradient-based optimization (targeting specific loss functions that maximize reasoning divergence while minimizing answer loss) to induce "Right Answer, Wrong Reasoning" behaviors. This vulnerability manifests through two primary attack vectors: 1. Token-level perturbations…

Robust Answers, Fragile Logic: Probing the Decoupling Hypothesis in LLM Reasoning
Evaluated models: Llama 3 8B, Mistral 7B, Zephyr 7B Beta +4 more

Source: arXiv

Model layer
Prompt layer
Hallucination
Embedding

Research methodology

Entries summarize publicly available primary-source security research. Model names reflect only systems explicitly evaluated by the cited paper, and measurements are research-reported unless independent verification is stated.