mlsanityData to 7 Oct 2026

Jailbreak susceptibility prediction and mitigation via the behavioral geometry of models

arXiv 2605.26409 ↗First posted 26 May 2026cs.CR cs.AI cs.LGcited by 0

On the map: LLM safety and security → Jailbreaking

Abstract (arXiv)

Evaluating and mitigating a generative system's susceptibility to jailbreak attacks is critical to its safe deployment. Given the number of deployable systems, full per-configuration evaluation and optimization is impractical. In this paper, we formalize the behavioral geometry of a population of models that, by leveraging previously evaluated and defended models, supports both efficient susceptibility prediction and effective defense transfer across a population. We apply the framework to 79 models spanning 24 providers and to 100 system configurations of a single base model. Simple methods that use the behavioral geometry reach an AUPRC of 0.94 for susceptibility detection with ≈98% fewer probes relative to a full evaluation. Using the behavioral geometry to select which model to transfer an optimized defense from outperforms same-provider assignment (+2%, p = 0.03) at no additional probe cost, with a set of three models sufficient to cover the population. Results are robust to hyperparameter selection and judge.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Jailbreaking, one of 12 directions in LLM safety and security. Builds on 35 of its 45 references that are in the map. First posted on arXiv 26 May 2026.

Builds on

35 of 45 references.

Closest ancestors · 12

Same direction, newest first

  1. 2026MultiBreak: A Scalable and Diverse Multi-turn Jailbreak Benchmark for Evaluating LLM Safety3 May
  2. 2024Best-of-N Jailbreaking4 Dec
  3. 2024Jailbreak Attacks and Defenses Against Large Language Models: A Survey5 Jul
  4. 2024Great, Now Write an Article About That: The Crescendo Multi-Turn LLM Jailbreak Attack2 Apr
  5. 2024JailbreakBench: An Open Robustness Benchmark for Jailbreaking Large Language Models28 Mar
  6. 2024AutoDefense: Multi-Agent LLM Defense against Jailbreak Attacks2 Mar
  7. 2024Intention Analysis Makes LLMs A Good Jailbreak Defender12 Jan
  8. 2023Jailbreaking Black Box Large Language Models in Twenty Queries12 Oct
  9. 2023Jailbreak and Guard Aligned Language Models with Only Few In-Context Demonstrations10 Oct
  10. 2023SmoothLLM: Defending Large Language Models Against Jailbreaking Attacks5 Oct
  11. 2023Defending Against Alignment-Breaking Attacks via Robustly Aligned LLM18 Sep
  12. 2023Jailbroken: How Does LLM Safety Training Fail?5 Jul

Other references · 23

  1. 2023GPT-4 Technical Reportcited by 18,132
  2. 2024The Llama 3 Herd of Modelscited by 14,192
  3. 2023Judging LLM-as-a-Judge with MT-Bench and Chatbot Arenacited by 8,393
  4. 2024GPT-4o System Cardcited by 4,467
  5. 2024Gemini 1.5: Unlocking multimodal understanding across millions of tokens of contextcited by 3,175
  6. 2024Gemma 2: Improving Open Language Models at a Practical Sizecited by 1,928
  7. 2019Similarity of Neural Network Representations Revisitedcited by 1,695
  8. 2024HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusalcited by 1,330
  9. 2023Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversationscited by 1,124
  10. 2022Ignore Previous Prompt: Attack Techniques For Language Modelscited by 760
  11. 2020Intrinsic Dimensionality Explains the Effectiveness of Language Model Fine-Tuningcited by 658
  12. 2024The Platonic Representation Hypothesiscited by 362
11 more
  1. 2024tinyBenchmarks: evaluating LLMs with fewer examplescited by 295
  2. 2024Nomic Embed: Training a Reproducible Long Context Text Embeddercited by 255
  3. 2023LLM Self Defense: By Self Examination, LLMs Know They Are Being Trickedcited by 226
  4. 2023Anchor Points: Benchmarking Models with Much Fewer Examplescited by 84
  5. 2023Comparing Foundation Models using Data Kernelscited by 14
  6. 2024Tracking the perspectives of interacting language modelscited by 14
  7. 2024Consistent estimation of generative model representations in the data kernel perspective spacecited by 10
  8. 2024Statistical inference on black-box generative models in the data kernel perspective spacecited by 9
  9. 2024Continuous Multidimensional Scalingcited by 4
  10. 2025Detecting Perspective Shifts in Multi-agent Systemscited by 4
  11. 2026Query-efficient model evaluation using cached responsescited by 2

Built on by

Not cited yet.