Step-by-Step Reasoning Attack: Revealing 'Erased' Knowledge in Large Language Models

TL;DR

This paper introduces a step-by-step reasoning-based black-box attack called Sleek that exposes failures in current knowledge erasure methods in large language models, revealing hidden or suppressed information.

Contribution

It presents a novel attack framework leveraging structured prompts to systematically recover erased knowledge in LLMs, exposing weaknesses in existing unlearning techniques.

Findings

62.5% success in retrieving forgotten Harry Potter facts

50% of prompts exposed unfair suppression of retained knowledge

Existing unlearning methods are vulnerable to structured reasoning attacks

Abstract

Knowledge erasure in large language models (LLMs) is important for ensuring compliance with data and AI regulations, safeguarding user privacy, mitigating bias, and misinformation. Existing unlearning methods aim to make the process of knowledge erasure more efficient and effective by removing specific knowledge while preserving overall model performance, especially for retained information. However, it has been observed that the unlearning techniques tend to suppress and leave the knowledge beneath the surface, thus making it retrievable with the right prompts. In this work, we demonstrate that \textit{step-by-step reasoning} can serve as a backdoor to recover this hidden information. We introduce a step-by-step reasoning-based black-box attack, Sleek, that systematically exposes unlearning failures. We employ a structured attack framework with three core components: (1) an adversarial prompt generation strategy leveraging step-by-step reasoning built from LLM-generated queries, (2) an attack mechanism that successfully recalls erased content, and exposes unfair suppression of knowledge intended for retention and (3) a categorization of prompts as direct, indirect, and implied, to identify which query types most effectively exploit unlearning weaknesses. Through extensive evaluations on four state-of-the-art unlearning techniques and two widely used LLMs, we show that existing approaches fail to ensure reliable knowledge removal. Of the generated adversarial prompts, 62.5% successfully retrieved forgotten Harry Potter facts from WHP-unlearned Llama, while 50% exposed unfair suppression of retained knowledge. Our work highlights the persistent risks of information leakage, emphasizing the need for more robust unlearning strategies for erasure.