PD3F: A Pluggable and Dynamic DoS-Defense Framework Against Resource Consumption Attacks Targeting Large Language Models
Not on the map. Why?
Abstract (arXiv)
Large Language Models (LLMs), due to substantial computational requirements, are vulnerable to resource consumption attacks, which can severely degrade server performance or even cause crashes, as demonstrated by denial-of-service (DoS) attacks designed for LLMs. However, existing works lack mitigation strategies against such threats, resulting in unresolved security risks for real-world LLM deployments. To this end, we propose the Pluggable and Dynamic DoS-Defense Framework (PD3F), which employs a two-stage approach to defend against resource consumption attacks from both the input and output sides. On the input side, we propose the Resource Index to guide Dynamic Request Polling Scheduling, thereby reducing resource usage induced by malicious attacks under high-concurrency scenarios. On the output side, we introduce the Adaptive End-Based Suppression mechanism, which terminates excessive malicious generation early. Experiments across six models demonstrate that PD3F significantly mitigates resource consumption attacks, improving users' access capacity by up to 500% during adversarial load. PD3F represents a step toward the resilient and resource-aware deployment of LLMs against resource consumption attacks.
Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Builds on 57 of its 88 references that are in the map. First posted on arXiv 24 May 2025.
Builds on
57 of 88 references.
References · 57
- 2017Attention Is All You Needcited by 60,361
- 2022Training language models to follow instructions with human feedbackcited by 15,971
- 2021Evaluating Large Language Models Trained on Codecited by 8,320
- 2020Measuring Massive Multitask Language Understandingcited by 7,963
- 2019HellaSwag: Can a Machine Really Finish Your Sentence?cited by 4,100
- 2023Universal and Transferable Adversarial Attacks on Aligned Language Modelscited by 3,104
- 2023GPQA: A Graduate-Level Google-Proof Q&A Benchmarkcited by 2,879
- 2023Jailbroken: How Does LLM Safety Training Fail?cited by 1,651
- 2023Jailbreaking Black Box Large Language Models in Twenty Queriescited by 1,420
- 2020RealToxicityPrompts: Evaluating Neural Toxic Degeneration in Language Modelscited by 1,255
- 2021Scaling Language Models: Methods, Analysis & Insights from Training Gophercited by 1,202
- 2022Ignore Previous Prompt: Attack Techniques For Language Modelscited by 832
38 more
- 2020Aligning AI With Shared Human Valuescited by 723
- 2023Tree of Attacks: Jailbreaking Black-Box LLMs Automaticallycited by 672
- 2023Safe RLHF: Safe Reinforcement Learning from Human Feedbackcited by 656
- 2023Baseline Defenses for Adversarial Attacks Against Aligned Language Modelscited by 651
- 2019Defending Against Neural Fake Newscited by 559
- 2023Jailbreaking ChatGPT via Prompt Engineering: An Empirical Studycited by 498
- 2023SmoothLLM: Defending Large Language Models Against Jailbreaking Attackscited by 469
- 2023Detecting Language Model Attacks with Perplexitycited by 431
- 2023Trustworthy LLMs: a Survey and Guideline for Evaluating Large Language Models' Alignmentcited by 401
- 2023Multi-step Jailbreaking Privacy Attacks on ChatGPTcited by 333
- 2023Poisoning Language Models During Instruction Tuningcited by 269
- 2023Certifying LLM Safety against Adversarial Promptingcited by 266
- 2023LLM Self Defense: By Self Examination, LLMs Know They Are Being Trickedcited by 244
- 2024AlphaEdit: Null-Space Constrained Knowledge Editing for Language Modelscited by 236
- 2023LoRA Fine-tuning Efficiently Undoes Safety Training in Llama 2-Chat 70Bcited by 195
- 2022The Carbon Footprint of Machine Learning Training Will Plateau, Then Shrinkcited by 184
- 2025Multi-agent Architecture Search via Agentic Supernetcited by 179
- 2024AmpleGCG: Learning a Universal and Transferable Generative Model of Adversarial Suffixes for Jailbreaking Both Open and Closed LLMscited by 172
- 2024Robust Prompt Optimization for Defending Language Models Against Jailbreaking Attackscited by 156
- 2024A Note on Shumailov et al. (2024): `AI Models Collapse When Trained on Recursively Generated Data'cited by 115
- 2023Explore, Establish, Exploit: Red Teaming Language Models from Scratchcited by 114
- 2020Sponge Examples: Energy-Latency Attacks on Neural Networkscited by 111
- 2025OverThink: Slowdown Attacks on Reasoning LLMscited by 96
- 2024Coercing LLMs to do and reveal (almost) anythingcited by 89
- 2025AnyEdit: Edit Any Knowledge Encoded in Language Modelscited by 76
- 2024Risk Taxonomy, Mitigation, and Assessment Benchmarks of Large Language Model Systemscited by 66
- 2024Speak Out of Turn: Safety Vulnerability of Large Language Models in Multi-turn Dialoguecited by 66
- 2024Pandora: Jailbreak GPTs by Retrieval Augmented Generation Poisoningcited by 64
- 2025Safety in Large Reasoning Models: A Surveycited by 64
- 2024Make Every Penny Count: Difficulty-Adaptive Self-Consistency for Cost-Efficient Reasoningcited by 63
- 2023Black Box Adversarial Prompting for Foundation Modelscited by 62
- 2023Guiding Language Model Reasoning with Planning Tokenscited by 58
- 2024Emulated Disalignment: Safety Alignment for Large Language Models May Backfire!cited by 37
- 2024Crabs: Consuming Resource via Auto-generation for LLM-DoS Attack under Black-box Settingscited by 35
- 2023Text-CRS: A Generalized Certified Robustness Framework against Textual Adversarial Attackscited by 34
- 2024Prompting Techniques for Secure Code Generation: A Systematic Investigationcited by 34
- 2024An Engorgio Prompt Makes Large Language Model Babble oncited by 30
- 2025SafeMLRM: Demystifying Safety in Multi-modal Large Reasoning Modelscited by 29
50 most cited of 57 All 57 (JSON)
Built on by
Citing papers per year: 2026: 1.
1 paper.