MirrorShield: Towards Universal Defense Against Jailbreaks via Entropy-Guided Mirror Crafting
On the map: LLM safety and security → Jailbreaking
Abstract (arXiv)
Defending large language models (LLMs) against jailbreak attacks is crucial for ensuring their safe deployment. Existing defense strategies typically rely on predefined static criteria to differentiate between harmful and benign prompts. However, such rigid rules fail to accommodate the inherent complexity and dynamic nature of real-world jailbreak attacks. In this paper, we focus on the novel challenge of universal defense against diverse jailbreaks. We propose a new concept ``mirror'', which is a dynamically generated prompt that reflects the syntactic structure of the input while ensuring semantic safety. The discrepancies between input prompts and their corresponding mirrors serve as guiding principles for defense. A novel defense model, MirrorShield, is further proposed to detect and calibrate risky inputs based on the crafted mirrors. Evaluated on multiple benchmark datasets and compared against ten state-of-the-art attack methods, MirrorShield demonstrates superior defense performance and promising generalization capabilities.
Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Placed in Jailbreaking, one of 12 directions in LLM safety and security. Builds on 30 of its 48 references that are in the map. First posted on arXiv 17 Mar 2025.
Builds on
30 of 48 references.
Closest ancestors · 17
Same direction, newest first
- 2025Layer-Level Self-Exposure and Patch: Affirmative Token Mitigation for Jailbreak Attack Defense5 Jan
- 2024RobustKV: Defending Large Language Models against Jailbreak Attacks via KV Eviction25 Oct
- 2024Root Defence Strategies: Ensuring Safety of LLM at the Decoding Level9 Oct
- 2024SafeAligner: Safety Alignment against Jailbreak Attacks via Response Disparity Guidance26 Jun
- 2024How Alignment and Jailbreak Work: Explain LLM Safety through Intermediate Hidden States9 Jun
- 2024Making Them Ask and Answer: Jailbreaking Large Language Models in Few Queries via Disguise and Reconstruction28 Feb
- 2024DrAttack: Prompt Decomposition and Reconstruction Makes Powerful LLM Jailbreakers25 Feb
- 2024SafeDecoding: Defending against Jailbreak Attacks via Safety-Aware Decoding14 Feb
- 2024Robust Prompt Optimization for Defending Language Models Against Jailbreaking Attacks30 Jan
- 2023Tree of Attacks: Jailbreaking Black-Box LLMs Automatically4 Dec
- 2023DeepInception: Hypnotize Large Language Model to Be Jailbreaker6 Nov
- 2023Self-Guard: Empower the LLM to Safeguard Itself24 Oct
5 more
- 2023Jailbreaking Black Box Large Language Models in Twenty Queries12 Oct
- 2023Jailbreak and Guard Aligned Language Models with Only Few In-Context Demonstrations10 Oct
- 2023SmoothLLM: Defending Large Language Models Against Jailbreaking Attacks5 Oct
- 2023GPTFUZZER: Red Teaming Large Language Models with Auto-Generated Jailbreak Prompts19 Sep
- 2023Detecting Language Model Attacks with Perplexity27 Aug
Other references · 13
- 2019Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformercited by 14,491
- 2023Llama 2: Open Foundation and Fine-Tuned Chat Modelscited by 12,865
- 2021Finetuned Language Models Are Zero-Shot Learnerscited by 3,747
- 2023Fine-tuning Aligned Language Models Compromises Safety, Even When Users Do Not Intend To!cited by 1,106
- 2023AlpacaFarm: A Simulation Framework for Methods that Learn from Human Feedbackcited by 738
- 2023Baseline Defenses for Adversarial Attacks Against Aligned Language Modelscited by 608
- 2021MetaICL: Learning to Learn In Contextcited by 450
- 2023LLM Self Defense: By Self Examination, LLMs Know They Are Being Trickedcited by 226
- 2024Attacks, Defenses and Evaluations for LLM Conversation Safety: A Surveycited by 124
- 2023Evil Geniuses: Delving into the Safety of LLM-based Agentscited by 103
- 2023Controlled Text Generation with Natural Language Instructionscited by 87
- 2022Entropy-based Attention Regularization Frees Unintended Bias Mitigation from Listscited by 54
1 more
Built on by
Citing papers per year: 2025: 1.
1 paper.