mlsanityData to 14 Oct 2026

Large Language Model Safety: A Holistic Survey

arXiv 2412.17686 ↗First posted 23 Dec 2024cs.AI cs.CLcited by 59

On the map: LLM safety and security → Safety alignment, refusal and guardrails

Abstract (arXiv)

The rapid development and deployment of large language models (LLMs) have introduced a new frontier in artificial intelligence, marked by unprecedented capabilities in natural language understanding and generation. However, the increasing integration of these models into critical applications raises substantial safety concerns, necessitating a thorough examination of their potential risks and associated mitigation strategies. This survey provides a comprehensive overview of the current landscape of LLM safety, covering four major categories: value misalignment, robustness to adversarial attacks, misuse, and autonomous AI risks. In addition to the comprehensive review of the mitigation methodologies and evaluation resources on these four aspects, we further explore four topics related to LLM safety: the safety implications of LLM agents, the role of interpretability in enhancing LLM safety, the technology roadmaps proposed and abided by a list of AI companies and institutes for LLM safety, and AI governance aimed at LLM safety with discussions on international cooperation, policy proposals, and prospective regulatory directions. Our findings underscore the necessity for a proactive, multifaceted approach to LLM safety, emphasizing the integration of technical solutions, ethical considerations, and robust governance frameworks. This survey is intended to serve as a foundational resource for academy researchers, industry practitioners, and policymakers, offering insights into the challenges and opportunities associated with the safe integration of LLMs into society. Ultimately, it seeks to contribute to the safe and beneficial development of LLMs, aligning with the overarching goal of harnessing AI for societal advancement and well-being. A curated list of related papers has been publicly available at https://github.com/tjunlp-lab/Awesome-LLM-Safety-Papers.

Key facts. Cited by 59 papers in the map (data through 14 Oct 2026). Placed in Safety alignment, refusal and guardrails, one of 12 directions in LLM safety and security. Builds on 228 of its 554 references that are in the map. First posted on arXiv 23 Dec 2024.

Builds on

228 of 554 references.

Closest ancestors · 8

Same direction, newest first

  1. 2024The Multilingual Alignment Prism: Aligning Global and Local Preferences to Reduce Harm26 Jun
  2. 2024WildGuard: Open One-Stop Moderation Tools for Safety Risks, Jailbreaks, and Refusals of LLMs26 Jun
  3. 2024Current state of LLM Risks and AI Guardrails16 Jun
  4. 2024The WMDP Benchmark: Measuring and Reducing Malicious Use With Unlearning5 Mar
  5. 2024Large Language Models are Vulnerable to Bait-and-Switch Attacks for Generating Harmful Content21 Feb
  6. 2023LoRA Fine-tuning Efficiently Undoes Safety Training in Llama 2-Chat 70B31 Oct
  7. 2023Self-Detoxifying Language Models via Toxification Reversal14 Oct
  8. 2023The Capacity for Moral Self-Correction in Large Language Models15 Feb

Other references · 220

  1. 2023GPT-4 Technical Reportcited by 18,396
  2. 2017Proximal Policy Optimization Algorithmscited by 15,215
  3. 2024The Llama 3 Herd of Modelscited by 14,715
  4. 2023Llama 2: Open Foundation and Fine-Tuned Chat Modelscited by 13,028
  5. 2021On the Opportunities and Risks of Foundation Modelscited by 4,033
  6. 2022Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedbackcited by 3,720
  7. 2023Qwen Technical Reportcited by 3,344
  8. 2023Mistral 7Bcited by 3,164
  9. 2023Universal and Transferable Adversarial Attacks on Aligned Language Modelscited by 3,104
  10. 2022Constitutional AI: Harmlessness from AI Feedbackcited by 3,050
  11. 2023Sparks of Artificial General Intelligence: Early experiments with GPT-4cited by 2,592
  12. 2024Qwen2 Technical Reportcited by 2,120
38 more
  1. 2015High-Dimensional Continuous Control Using Generalized Advantage Estimationcited by 2,026
  2. 2016Concrete Problems in AI Safetycited by 1,642
  3. 2024HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusalcited by 1,429
  4. 2023Jailbreaking Black Box Large Language Models in Twenty Queriescited by 1,420
  5. 2017Towards A Rigorous Science of Interpretable Machine Learningcited by 1,416
  6. 2023Representation Engineering: A Top-Down Approach to AI Transparencycited by 1,336
  7. 2023Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversationscited by 1,207
  8. 2024DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Modelcited by 1,135
  9. 2021A General Language Assistant as a Laboratory for Alignmentcited by 1,055
  10. 2022Toy Models of Superpositioncited by 971
  11. 2021Ethical and social risks of harm from Language Modelscited by 966
  12. 2022In-context Learning and Induction Headscited by 964
  13. 2023Gemini: A Family of Highly Capable Multimodal Modelscited by 872
  14. 2022Ignore Previous Prompt: Attack Techniques For Language Modelscited by 832
  15. 2022Quantifying Memorization Across Neural Language Modelscited by 763
  16. 2019Universal Adversarial Triggers for Attacking and Analyzing NLPcited by 722
  17. 2024Yi: Open Foundation Models by 01.AIcited by 719
  18. 2023Tree of Attacks: Jailbreaking Black-Box LLMs Automaticallycited by 672
  19. 2023Baichuan 2: Open Large-scale Language Modelscited by 670
  20. 2024A Survey on Large Language Models for Code Generationcited by 658
  21. 2023Bias and Fairness in Large Language Models: A Surveycited by 639
  22. 2022Grokking: Generalization Beyond Overfitting on Small Algorithmic Datasetscited by 629
  23. 2023Prompt Injection attack against LLM-integrated Applicationscited by 605
  24. 2024How Johnny Can Persuade LLMs to Jailbreak Them: Rethinking Persuasion to Challenge AI Safety by Humanizing LLMscited by 573
  25. 2023GPTFUZZER: Red Teaming Large Language Models with Auto-Generated Jailbreak Promptscited by 560
  26. 2024Scaling and evaluating sparse autoencoderscited by 553
  27. 2023Jailbreaking ChatGPT via Prompt Engineering: An Empirical Studycited by 498
  28. 2023SmoothLLM: Defending Large Language Models Against Jailbreaking Attackscited by 469
  29. 2024Sparse Feature Circuits: Discovering and Editing Interpretable Causal Graphs in Language Modelscited by 435
  30. 2023Scalable Extraction of Training Data from (Production) Language Modelscited by 434
  31. 2023Detecting Language Model Attacks with Perplexitycited by 431
  32. 2023Jailbreak and Guard Aligned Language Models with Only Few In-Context Demonstrationscited by 431
  33. 2024Mechanistic Interpretability for AI Safety -- A Reviewcited by 422
  34. 2021BOLD: Dataset and Metrics for Measuring Biases in Open-Ended Language Generationcited by 412
  35. 2021Large Language Models Can Be Strong Differentially Private Learnerscited by 412
  36. 2023Trustworthy LLMs: a Survey and Guideline for Evaluating Large Language Models' Alignmentcited by 401
  37. 2023Who's Harry Potter? Approximate Unlearning in LLMscited by 381
  38. 2023Orca: Progressive Learning from Complex Explanation Traces of GPT-4cited by 379

50 most cited of 220 All 220 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 34; 2026: 25.

59 papers.

Sort:

By area:

  1. 2025CollabLLM: From Passive Responders to Active Collaboratorscited by 124
  2. 2025A Comprehensive Survey in LLM(-Agent) Full Stack Safety: Data, Training and Deploymentcited by 116
  3. 2025Safety in Large Reasoning Models: A Surveycited by 64
  4. 2025MCPSecBench: A Systematic Security Benchmark and Playground for Testing Model Context Protocolscited by 51
  5. 2025AlphaSteer: Learning Refusal Steering with Principled Null-Space Constraintcited by 49
  6. 2025The State of Multilingual LLM Safety Research: From Measuring the Language Gap to Mitigating Itcited by 45
  7. 2025We Should Identify and Mitigate Third-Party Safety Risks in MCP-Powered Agent Systemscited by 27
  8. 2025A Comprehensive Survey on Trustworthiness in Reasoning with Large Language Modelscited by 25
  9. 2025SAFEPATH: Preventing Harmful Reasoning in Chain-of-Thought via Early Alignmentcited by 24
  10. 2025A Survey on the Safety and Security Threats of Computer-Using Agents: JARVIS or Ultron?cited by 23
  11. 2025MTR-DuplexBench: Towards a Comprehensive Evaluation of Multi-Round Conversations for Full-Duplex Speech Language Modelscited by 22
  12. 2025Revisiting Entropy in Reinforcement Learning for Large Reasoning Modelscited by 21
38 more
  1. 2025False Sense of Security: Why Probing-based Malicious Input Detection Fails to Generalizecited by 19
  2. 2025LLM Unlearning with LLM Beliefscited by 17
  3. 2025Persona-Augmented Benchmarking: Evaluating LLMs Across Diverse Writing Stylescited by 16
  4. 2025When Refusals Fail: Unstable Safety Mechanisms in Long-Context LLM Agentscited by 15
  5. 2025Standard Benchmarks Fail -- Auditing LLM Agents in Finance Must Prioritize Riskcited by 13
  6. 2025Survey on AI-Generated Media Detection: From Non-MLLM to MLLMcited by 12
  7. 2025Building Safe GenAI Applications: An End-to-End Overview of Red Teaming for Large Language Modelscited by 11
  8. 2025Interpretation Meets Safety: A Survey on Interpretation Methods and Tools for Improving LLM Safetycited by 10
  9. 2025SafeVid: Toward Safety Aligned Video Large Multimodal Modelscited by 9
  10. 2025DecEx-RAG: Boosting Agentic Retrieval-Augmented Generation with Decision and Execution Optimization via Process Supervisioncited by 9
  11. 2025Beyond Chains of Thought: Benchmarking Latent-Space Reasoning Abilities in Large Language Modelscited by 6
  12. 2025A Survey of Foundation Models for IoT: Taxonomy and Criteria-Based Analysiscited by 5
  13. 2026State Contamination in Memory-Augmented LLM Agentscited by 5
  14. 2025Are Bias Evaluation Methods Biased ?cited by 4
  15. 2025SAFENLIDB: A Privacy-Preserving Safety Alignment Framework for LLM-based Natural Language Database Interfacescited by 4
  16. 2026FormalJudge: A Neuro-Symbolic Paradigm for Agentic Oversightcited by 4
  17. 2026Asymmetric Actor-Critic for Multi-turn LLM Agentscited by 4
  18. 2025Beyond Over-Refusal: Scenario-Based Diagnostics and Post-Hoc Mitigation for Exaggerated Refusals in LLMscited by 3
  19. 2025Surfacing Subtle Stereotypes: A Multilingual, Debate-Oriented Evaluation of Modern LLMscited by 3
  20. 2025MAS-Shield: A Defense Framework for Secure and Efficient LLM MAScited by 3
  21. 2026Why Does Reinforcement Learning Generalize? A Feature-Level Mechanistic Study of Post-Training in Large Language Modelscited by 3
  22. 2026Mitigating Memorization Where It Happenscited by 2
  23. 2026KUDA: Knowledge Unlearning by Deviating Representation for Large Language Modelscited by 2
  24. 2026ALTER: Asymmetric LoRA for Token-Entropy-Guided Unlearning of LLMscited by 2
  25. 2026An Empirical Study of Multi-Generation Sampling for Jailbreak Detection in Large Language Modelscited by 2
  26. 2026Layered LLM Defenses as an Ensemble: Access Tiers, Inference Cost, and the Measured Failure Correlation Between Defense Layerscited by 2
  27. 2026Analyzing Bias in False Refusal Behavior of Large Language Models for Hate Speech Detoxificationcited by 1
  28. 2026Preserving Fairness and Safety in Quantized LLMs Through Critical Weight Protectioncited by 1
  29. 2026Step-Wise Refusal Dynamics in Autoregressive and Diffusion Language Modelscited by 1
  30. 2026TrailBlazer: History-Guided Reinforcement Learning for Black-Box LLM Jailbreakingcited by 1
  31. 2026Common Sense vs. Morality: The Curious Case of Narrative Focus Bias in LLMscited by 1
  32. 2026JUBAKU: An Adversarial Benchmark for Exposing Culturally Grounded Stereotypes in Japanese LLMscited by 1
  33. 2026LLM-Guided Prompt Evolution for Password Guessingcited by 1
  34. 2025Certifiable Safe RLHF: Semantic Grounding and Fixed Penalty Constraint Optimization for Safer LLM Alignmentcited by 0
  35. 2025Validity Is What You Needcited by 0
  36. 2025Mirror, Mirror on the Wall -- Which is the Best Model of Them All?cited by 0
  37. 2025AIAuditTrack: A Framework for AI Security systemcited by 0
  38. 2025Constrained Language Model Policy Optimization via Risk-aware Stepwise Alignmentcited by 0

50 most cited of 59 All 59 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2025A Survey of Attacks on Large Language Models
  2. 2023Survey of Vulnerabilities in Large Language Models Revealed by Adversarial Attacks
  3. 2024ALERT: A Comprehensive Benchmark for Assessing Large Language Models' Safety through Red Teaming
  4. 2025Safety at Scale: A Comprehensive Survey of Large Model and Agent Safety
  5. 2024Detectors for Safe and Reliable LLMs: Implementations, Uses, and Limitations
  6. 2024Breaking Down the Defenses: A Comparative Survey of Attacks on Large Language Models
  7. 2025Security Concerns for Large Language Models: A Survey
  8. 2024AEGIS: Online Adaptive AI Content Safety Moderation with Ensemble of LLM Experts