mlsanityData to 14 Oct 2026

SHIELD: Classifier-Guided Prompting for Robust and Safer LVLMs

arXiv 2510.13190 ↗First posted 15 Oct 2025cs.CLcited by 5

On the map: LLM reasoning → Reasoning for agents, planning and code

also in 1 more
Abstract (arXiv)

Large Vision-Language Models (LVLMs) unlock powerful multimodal reasoning but also expand the attack surface, particularly through adversarial inputs that conceal harmful goals in benign prompts. We propose SHIELD, a lightweight, model-agnostic preprocessing framework that couples fine-grained safety classification with category-specific guidance and explicit actions (Block, Reframe, Forward). Unlike binary moderators, SHIELD composes tailored safety prompts that enforce nuanced refusals or safe redirection without retraining. Across five benchmarks and five representative LVLMs, SHIELD consistently lowers jailbreak and non-following rates while preserving utility. Our method is plug-and-play, incurs negligible overhead, and is easily extendable to new attack types -- serving as a practical safety patch for both weakly and strongly aligned LVLMs.

Key facts. Cited by 5 papers in the map (data through 14 Oct 2026). Placed in Reasoning for agents, planning and code, one of 17 directions in LLM reasoning. Builds on 21 of its 35 references that are in the map. First posted on arXiv 15 Oct 2025.

Builds on

21 of 35 references.

Closest ancestors · 1

Same direction, newest first

  1. 2024Deliberative Alignment: Reasoning Enables Safer Language Models20 Dec

Other references · 20

  1. 2022Training language models to follow instructions with human feedbackcited by 15,971
  2. 2017Proximal Policy Optimization Algorithmscited by 15,215
  3. 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 8,065
  4. 2023Universal and Transferable Adversarial Attacks on Aligned Language Modelscited by 3,104
  5. 2023DSPy: Compiling Declarative Language Model Calls into Self-Improving Pipelinescited by 832
  6. 2024Safety Alignment Should Be Made More Than Just a Few Tokens Deepcited by 472
  7. 2023FigStep: Jailbreaking Large Vision-Language Models via Typographic Visual Promptscited by 398
  8. 2023Jailbreak in pieces: Compositional Adversarial Attacks on Multi-Modal Language Modelscited by 273
  9. 2024JailBreakV: A Benchmark for Assessing the Robustness of MultiModal Large Language Models against Jailbreak Attackscited by 230
  10. 2024Llama Guard 3 Vision: Safeguarding Human-AI Image Understanding Conversationscited by 129
  11. 2024AdaShield: Safeguarding Multimodal Large Language Models from Structure-based Attack via Adaptive Shield Promptingcited by 125
  12. 2024ETA: Evaluating Then Aligning Safety of Vision Language Models at Inference Timecited by 60
8 more
  1. 2024LlavaGuard: An Open VLM-based Framework for Safeguarding Vision Datasets and Modelscited by 52
  2. 2024Steering Away from Harm: An Adaptive Approach to Defending Vision Language Model Against Jailbreakscited by 48
  3. 2024PAD: Patch-Agnostic Defense against Adversarial Patch Attackscited by 32
  4. 2024BlueSuffix: Reinforced Blue Teaming for Vision-Language Models Against Jailbreak Attackscited by 28
  5. 2024Safeguarding Vision-Language Models Against Patched Visual Prompt Injectorscited by 27
  6. 2024The VLLM Safety Paradox: Dual Ease in Jailbreak Attack and Defensecited by 21
  7. 2025FC-Attack: Jailbreaking Multimodal Large Language Models via Auto-Generated Flowchartscited by 18
  8. 2025Seeing the Threat: Vulnerabilities in Vision-Language Models to Adversarial Attackcited by 5

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 1; 2026: 4.

5 papers.

  1. 2026Physical Prompt Injection Attacks on Large Vision-Language Modelscited by 4
  2. 2025Agentic Moderation: Multi-Agent Design for Safer Vision-Language Modelscited by 2
  3. 2026When the Model Said 'No Comment', We Knew Helpfulness Was Dead, Honesty Was Alive, and Safety Was Terrifiedcited by 1
  4. 2026Do Clinical Question Answering Systems Really Need Specialised Medical Fine Tuning?cited by 0
  5. 2026Structure-Guided Visual Perturbation Neutralization for LVLMscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025VLMGuard-R1: Proactive Safety Alignment for VLMs via Reasoning-Driven Prompt Optimization
  2. 2025Spot Risks Before Speaking! Unraveling Safety Attention Heads in Large Vision-Language Models
  3. 2024RapGuard: Safeguarding Multimodal Large Language Models via Rationale-aware Defensive Prompting
  4. 2023Robust Safety Classifier for Large Language Models: Adversarial Prompt Shield
  5. 2026Reflex-Guard: A Low-Latency Guardrail for LLM Prompt Safety Using Dense Semantic Embeddings
  6. 2026SafeCap: Improving LVLM Safety with Image Captioning Reinforcement Learning
  7. 2025SGuard-v1: Safety Guardrail for Large Language Models
  8. 2026Reflect-Guard: Enhancing LLM Safeguards against Adversarial Prompts via Logical Self-Reflection