Updated 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
Affects: Llama 3 8B, Mistral 7B, Zephyr 7B Beta +4 more
Source: arXiv