mlsanityData to 7 Oct 2026

HumorReject: Decoupling LLM Safety from Refusal Prefix via A Little Humor

arXiv 2501.13677 ↗First posted 23 Jan 2025cs.LG cs.CRcited by 5

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

Abstract (arXiv)

Large Language Models (LLMs) commonly rely on explicit refusal prefixes for safety, making them vulnerable to prefix injection attacks. We introduce HumorReject, a novel data-driven approach that reimagines LLM safety by decoupling it from refusal prefixes through humor as an indirect refusal strategy. Rather than explicitly rejecting harmful instructions, HumorReject responds with contextually appropriate humor that naturally defuses potentially dangerous requests. Our approach effectively addresses common "over-defense" issues while demonstrating superior robustness against various attack vectors. Our findings suggest that improvements in training data design can be as important as the alignment algorithm itself in achieving effective LLM safety. The code and dataset are available at https://github.com/wooozihui/HumorReject.

Key facts. Cited by 5 papers in the map (data through 7 Oct 2026). Placed in Safety alignment, refusal and guardrails, one of 12 directions in LLM safety and security. Builds on 32 of its 46 references that are in the map. First posted on arXiv 23 Jan 2025.

Builds on

32 of 46 references.

Closest ancestors · 5

Same direction, newest first

  1. 2024Refusal in Language Models Is Mediated by a Single Direction17 Jun
  2. 2024Safety Alignment Should Be Made More Than Just a Few Tokens Deep10 Jun
  3. 2024Improving Alignment and Robustness with Circuit Breakers6 Jun
  4. 2024CodeAttack: Revealing Safety Generalization Challenges of Large Language Models via Code Completion12 Mar
  5. 2023Fine-tuning Aligned Language Models Compromises Safety, Even When Users Do Not Intend To!5 Oct

Other references · 27

  1. 2022Training language models to follow instructions with human feedbackcited by 15,712
  2. 2024The Llama 3 Herd of Modelscited by 14,192
  3. 2021LoRA: Low-Rank Adaptation of Large Language Modelscited by 14,084
  4. 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 7,916
  5. 2020Measuring Massive Multitask Language Understandingcited by 7,768
  6. 2017Deep reinforcement learning from human preferencescited by 3,846
  7. 2023Mistral 7Bcited by 3,065
  8. 2023Universal and Transferable Adversarial Attacks on Aligned Language Modelscited by 2,914
  9. 2024Gemma 2: Improving Open Language Models at a Practical Sizecited by 1,928
  10. 2023Jailbroken: How Does LLM Safety Training Fail?cited by 1,565
  11. 2024LlamaFactory: Unified Efficient Fine-Tuning of 100+ Language Modelscited by 1,442
  12. 2024HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusalcited by 1,330
15 more
  1. 2023Jailbreaking Black Box Large Language Models in Twenty Queriescited by 1,326
  2. 2023AutoDAN: Generating Stealthy Jailbreak Prompts on Aligned Large Language Modelscited by 738
  3. 2024ORPO: Monolithic Preference Optimization without Reference Modelcited by 556
  4. 2020On Adaptive Attacks to Adversarial Example Defensescited by 547
  5. 2024How Johnny Can Persuade LLMs to Jailbreak Them: Rethinking Persuasion to Challenge AI Safety by Humanizing LLMscited by 533
  6. 2023XSTest: A Test Suite for Identifying Exaggerated Safety Behaviours in Large Language Modelscited by 477
  7. 2024Jailbreaking Leading Safety-Aligned LLMs with Simple Adaptive Attackscited by 450
  8. 2023A Wolf in Sheep's Clothing: Generalized Nested Jailbreak Prompts can Fool Large Language Models Easilycited by 242
  9. 2023Multilingual Jailbreak Challenges in Large Language Modelscited by 222
  10. 2024Refuse Whenever You Feel Unsafe: Improving Safety in LLMs via Decoupled Refusal Trainingcited by 67
  11. 2024CodeChameleon: Personalized Encryption Framework for Jailbreaking Large Language Modelscited by 65
  12. 2023Let's Think Outside the Box: Exploring Leap-of-Thought in Large Language Models with Creative Humor Generationcited by 56
  13. 2024The Dark Side of Function Calling: Pathways to Jailbreaking Large Language Modelscited by 20
  14. 2024Humor Mechanics: Advancing Humor Generation with Multistep Reasoningcited by 8
  15. 2024CleanComedy: Creating Friendly Humor through Generative Techniquescited by 5

Built on by

Cited by year (2026 to 7 Oct)

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

5 papers.

  1. 2025Large Language Models for Subjective Language Understanding: A Surveycited by 9
  2. 2026TraceRouter: Robust Safety for Large Foundation Models via Path-Level Interventioncited by 8
  3. 2026Same Model, Different Weakness: How Language and Modality Reshape the Jailbreak Attack Surface in Frontier MLLMscited by 0
  4. 2026Incomplete Prompt Jailbreaks in Large Language Modelscited by 0
  5. 2026Towards High-Level Semantic Intelligencecited by 0