An LLM can Fool Itself: A Prompt-Based Adversarial Attack

Xilie Xu, Keyi Kong, Ning Liu, Li-zhen Cui, Di Wang +2 more

2/12/2026

No official code found. Best implementation: GodXuxilie/PromptAttack (114 stars, last commit 2025-01-21).

Semantic Scholar

Code Implementations(1)

GodXuxilie/PromptAttackCommunity67%

An LLM can Fool Itself: A Prompt-Based Adversarial Attack (ICLR 2024)

11417Python, Jupyter NotebookSep 19, 20231 years ago

Similar Papers

A Good Learner can Teach Better: Teacher-Student Collaborative Knowledge Distillation

100% match1 repo2026

Everything Everywhere All at Once: LLMs can In-Context Learn Multiple Tasks in Superposition

Context-PIPs: Persistent Independent Particles Demands Context Features

Uni-Mol: A Universal 3D Molecular Representation Learning Framework

RH-BrainFS: Regional Heterogeneous Multimodal Brain Networks Fusion Strategy

81% match1 repo72026

MS-TIP: Imputation Aware Pedestrian Trajectory Prediction

81% match1 repo242026

Cite this paper

@article{xu2026fool,
  title  = {An LLM can Fool Itself: A Prompt-Based Adversarial Attack},
  author = {Xilie Xu and Keyi Kong and Ning Liu and Li-zhen Cui and Di Wang and Jingfeng Zhang and Mohan S. Kankanhalli},
  year   = {2026},
  url    = {https://api.semanticscholar.org/CorpusID:3fc7fc724b2332382398fadb8dc7863eb43f7c93},
  journal = {ICLR 2024 2024}
}

Add a badge to your README

Shows whether official code is linked for this paper.

[![Code of Paper](https://codeofpaper.com/badge/3fc7fc724b2332382398fadb8dc7863eb43f7c93.svg)](https://codeofpaper.com/paper/3fc7fc724b2332382398fadb8dc7863eb43f7c93)

Discussion