JevAdvBench: A Benchmark and Black-Box Attacks for Reinforcement Learning for Calibrated Decisions Models

arXiv:2609.31142 · cs.CR, cs.AI, cs.CL · Submitted 2026-09-25 · Read on arXiv

cs.CR, cs.AI, cs.CL

Submitted: 2026-09-25

Updated: 2026-09-25

Comments: 33 pages, 13 figures, 19 tables. Project website: https://JevAdvBench.github.io/JevAdvBench/

Project page: https://jevadvbench.github.io/JevAdvBench/ABSTRACT

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Benchmark — 27× in this paper · explained in 2 episodes
• Robustness — 21× in this paper · explained in 5 episodes
• Accuracy — 11× in this paper · explained in 1 episode
• Calibration — 6× in this paper · explained in 9 episodes
• Consistency — 5× in this paper · explained in 3 episodes
• LLM-as-a-Judge — 5× in this paper · explained in 3 episodes
• Reinforcement Learning — 5× in this paper · explained in 4 episodes
• AUROC — 3× in this paper · explained in 3 episodes
• Adversarial Attacks — 3× in this paper
• Alignment — 3× in this paper
• Benchmarking — 3× in this paper · explained in 4 episodes
• Decode — 3× in this paper · explained in 2 episodes
• Provenance — 3× in this paper · explained in 5 episodes
• Sycophancy — 3× in this paper · explained in 5 episodes
• Verification — 3× in this paper · explained in 2 episodes
• LLM judge — 2× in this paper · explained in 2 episodes
• SIGSAC Conference on Computer and Communications Security — 2× in this paper
• Steering — 2× in this paper · explained in 2 episodes
• Trace — 2× in this paper · explained in 2 episodes
• AI Alignment — 1× in this paper · explained in 2 episodes
• Abstention — 1× in this paper · explained in 1 episode
• Adaptivity — 1× in this paper · explained in 1 episode
• Adversarial Examples — 1× in this paper · explained in 2 episodes
• Adversarial Training — 1× in this paper
• Chain-of-Thought Prompting — 1× in this paper · explained in 2 episodes
• Clustering — 1× in this paper · explained in 2 episodes
• Data Augmentation — 1× in this paper · explained in 7 episodes
• Grounding — 1× in this paper · explained in 3 episodes
• Hysteresis — 1× in this paper · explained in 3 episodes

Sources

• JailbreakBench: An Open Robustness Benchmark for Jailbreaking Large Language Models
• StruQ: Defending Against Prompt Injection with Structured Queries
• SecAlign: Defending Against Prompt Injection with Preference Optimization
• On Calibration of Modern Neural Networks
• Just Ask Jev: Reinforcement Learning for Calibrated Decisions as a Zero-Shot Detector of AI Alignment Failures
• A Baseline for Detecting Misclassified and Out-of-Distribution Examples in Neural Networks
• Is BERT Really Robust? A Strong Baseline for Natural Language Attack on Text Classification and Entailment
• Language Models (Mostly) Know What They Know
• Formalizing and Benchmarking Prompt Injection Attacks and Defenses
• Ignore Previous Prompt: Attack Techniques For Language Models
• Type-Safe Is Not Error-Free: Typed Decision Models Follow the Option Name, Not the Definition Bound to It
• The Instruction Hierarchy: Training LLMs to Prioritize Privileged Instructions
• Benchmarking and Defending Against Indirect Prompt Injection Attacks on Large Language Models
• Calibrate Before Use: Improving Few-Shot Performance of Language Models
• One Token to Fool LLM-as-a-Judge
• Large Language Models Are Not Robust Multiple Choice Selectors
• PromptRobust: Towards Evaluating the Robustness of Large Language Models on Adversarial Prompts
• Universal and Transferable Adversarial Attacks on Aligned Language Models

Related papers

• SoK: AI-Augmented Binary Reversing
• Relaxed Sender Anonymity for CBDC Interbank Settlement: A Zero-Knowledge Approach on Permissioned EVM
• Calibration-Family Overfit: Why Trusted Sabotage Monitors Don't Transfer Across Lineages
• Estimating the Social Cost of Corporate Data Breaches
• Efficient Fuzzy PSI under One-Sided Assumptions
• Sealing the Audit-Runtime Gap for LLM Skills