arXiv AI June 30, 2026 By Bu\u{g}ra Alperen Ulu{\i}rmak, Rifat Kurban

EvalSafetyGap: A Hybrid Survey and Conceptual Framework for LLM Evaluation-Safety Failures

arXiv:2606. 30219v1 Announce Type: new Abstract: LLM evaluation and AI safety face a shared measurement problem: benchmark scores, reward-model signals, and reported safety metrics can improve while the latent properties they are meant to represent remain difficult to verify.

Summary generated by The Flow from the publisher's feed. The full article lives at arXiv AI.

arXiv AI

Jun 11

Models That Know How Evaluations Are Designed Score Safer

arXiv:2605. 28591v2 Announce Type: replace-cross Abstract: The validity of AI safety evaluations depends on models behaving consistently across controlled and deployment settings.

By Katharina Deckenbach, Haritz Puerto, Jonas Geiping, Sahar Abdelnabi

fine-tuning benchmarks safety

arXiv AI

Jul 31

Adversarial Pragmatics for AI Safety Evaluation: A Diagnostic Framework and Seed Benchmark for Language-Mediated Control

arXiv:2607. 01153v3 Announce Type: replace-cross Abstract: Safety evaluations for language models increasingly depend on judgments about ambiguous natural-language behaviour: whether a model followed an instruction, refused appropriately, complied with a policy, or misreported progress in an agentic task.

By Brett Reynolds

llms agents benchmarks safety

arXiv AI

Jul 2

Adversarial Pragmatics for AI Safety Evaluation: A Benchmark for Instruction Conflict, Embedded Commands, and Policy Ambiguity

arXiv:2607. 01153v1 Announce Type: cross Abstract: Safety evaluations for language models increasingly depend on judgments about ambiguous natural-language behaviour: whether a model has followed an instruction, refused appropriately, complied with a policy, resisted an embedded command, or misreported progress in an agentic task.

By Brett Reynolds

llms agents benchmarks safety

arXiv AI

Jul 9

Beyond Attack-Success Rate: Action-Graded Severity Scale for Tool-Using AI Agents

arXiv:2607. 07474v1 Announce Type: cross Abstract: Agentic red-teaming benchmarks report whether an injected agent was compromised as a single bit: the attack succeeded, or it did not.

By Harry Owiredu-Ashley

agents benchmarks safety

arXiv AI

Jul 28

Epistemic Norms for AI Safety and Alignment Research

arXiv:2607. 24243v1 Announce Type: new Abstract: Mainstream AI research emphasises capability growth and tolerates low failure rates when average-case performance is high.

By Keivan Navaie

safety

arXiv Machine Learning

Jun 30

The Alignment Auditor: A Bayesian Framework for Verifying and Refining LLM Objectives

arXiv:2510. 06096v3 Announce Type: replace Abstract: The objectives that Large Language Models (LLMs) implicitly optimize remain dangerously opaque, making trustworthy alignment and auditing a grand challenge.

By Matthieu Bou, Nyal Patel, Arjun Jagota, Satyapriya Krishna, Sonali Parbhoo

llms reinforcement-learning safety

Related stories

Models That Know How Evaluations Are Designed Score Safer

Adversarial Pragmatics for AI Safety Evaluation: A Diagnostic Framework and Seed Benchmark for Language-Mediated Control

Adversarial Pragmatics for AI Safety Evaluation: A Benchmark for Instruction Conflict, Embedded Commands, and Policy Ambiguity

Beyond Attack-Success Rate: Action-Graded Severity Scale for Tool-Using AI Agents

Epistemic Norms for AI Safety and Alignment Research

The Alignment Auditor: A Bayesian Framework for Verifying and Refining LLM Objectives