mlsanityData to 14 Oct 2026

Latent Personality Alignment: Improving Harmlessness Without Mentioning Harms

arXiv 2605.08496 ↗First posted 8 May 2026cs.AIcited by 1

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

Abstract (arXiv)

Current adversarial robustness methods for large language models require extensive datasets of harmful prompts (thousands to hundreds of thousands of examples), yet remain vulnerable to novel attack vectors and distributional shifts. We propose Latent Personality Alignment (LPA), a sample-efficient defense that achieves robustness by training models on abstract personality traits rather than specific harmful behaviors. Using fewer than 100 trait statements and latent adversarial training, LPA achieves comparable attack success rates to methods trained on 150k+ examples, while maintaining superior utility. Critically, LPA generalizes better to unseen attack distributions, reducing misclassification rates by 2.6x compared to baseline across six harm benchmarks -- without ever seeing harmful examples during training. Our results demonstrate that personality-based alignment offers a principled approach to building robust defenses with minimal cost.

Key facts. Cited by 1 paper 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 25 of its 28 references that are in the map. First posted on arXiv 8 May 2026.

Builds on

25 of 28 references.

Closest ancestors · 3

Same direction, newest first

  1. 2024OR-Bench: An Over-Refusal Benchmark for Large Language Models31 May
  2. 2023Fine-tuning Aligned Language Models Compromises Safety, Even When Users Do Not Intend To!5 Oct
  3. 2023Fundamental Limitations of Alignment in Large Language Models19 Apr

Other references · 22

  1. 2022Training language models to follow instructions with human feedbackcited by 15,971
  2. 2017Deep reinforcement learning from human preferencescited by 3,912
  3. 2023Universal and Transferable Adversarial Attacks on Aligned Language Modelscited by 3,104
  4. 2022Constitutional AI: Harmlessness from AI Feedbackcited by 3,050
  5. 2024HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusalcited by 1,429
  6. 2021A General Language Assistant as a Laboratory for Alignmentcited by 1,055
  7. 2022Red Teaming Language Models with Language Modelscited by 1,026
  8. 2022Discovering Language Model Behaviors with Model-Written Evaluationscited by 836
  9. 2020Aligning AI With Shared Human Valuescited by 723
  10. 2024JailbreakBench: An Open Robustness Benchmark for Jailbreaking Large Language Modelscited by 536
  11. 2023Jailbreak and Guard Aligned Language Models with Only Few In-Context Demonstrationscited by 431
  12. 2023Low-Resource Languages Jailbreak GPT-4cited by 309
10 more
  1. 2023Do-Not-Answer: A Dataset for Evaluating Safeguards in LLMscited by 184
  2. 2025Constitutional Classifiers: Defending against Universal Jailbreaks across Thousands of Hours of Red Teamingcited by 181
  3. 2024Latent Adversarial Training Improves Robustness to Persistent Harmful Behaviors in LLMscited by 173
  4. 2023Personality Traits in Large Language Modelscited by 169
  5. 2024Efficient Adversarial Training in LLMs with Continuous Attackscited by 113
  6. 2024Defending Against Unforeseen Failure Modes with Latent Adversarial Trainingcited by 88
  7. 2025Reward Shaping to Mitigate Reward Hacking in RLHFcited by 88
  8. 2024Robust LLM safeguarding via refusal feature adversarial trainingcited by 64
  9. 2025Persistent Instability in LLM's Personality Measurements: Effects of Scale, Reasoning, and Conversation Historycited by 26
  10. 2025Bullying the Machine: How Personas Increase LLM Vulnerabilitycited by 9

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026Imprinting: Knowledge Distillation Can Shift Student Behaviorcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026Efficient Safety Alignment of Language Models via Latent Personality Traits
  2. 2025Adversarial Alignment for LLMs Requires Simpler, Reproducible, and More Measurable Objectives
  3. 2025Probing the Robustness of Large Language Models Safety to Latent Perturbations
  4. 2026Safety Is Not Universal: The Selective Safety Trap in LLM Alignment
  5. 2025LatentGuard: Controllable Latent Steering for Robust Refusal of Attacks and Reliable Response Generation
  6. 2025In-Training Defenses against Emergent Misalignment in Language Models
  7. 2026Target-free Latent Safety Alignment
  8. 2025AdversariaLLM: A Unified and Modular Toolbox for LLM Robustness Research