Beyond Accuracy: Policy Invariance as a Reliability Test for LLM Safety Judges
LLM-as-a-Judge pipelines have become the de facto evaluator for agent safety, yet existing benchmarks treat their verdicts as ground-truth proxies without checking whether the verdicts depend on the agent's behavior or merely on how the evaluation policy happens to be worded. We argue that any trustworthy safety judge must satisfy a basic property we call policy invariance, and we operationalize it as three testable principles: rubric-semantics invariance under certified-equivalent rewrites, rubric-threshold invariance under intentional strict-to-lenient shifts, and ambiguity-aware calibration so that verdict instability concentrates on genuinely ambiguous cases. Instantiating these principles as a stress-test protocol with four agent-class judges on trajectories drawn from ASSEBench and R-Judge, we surface a previously unmeasured failure mode: today's judges respond to meaningful normative shifts and to meaningless structural rewrites with comparable strength, and cannot tell the two apart. Content-preserving policy rewrites flip up to 9.1% of verdicts above baseline jitter, and 18-43% of all observed flips occur on unambiguous cases under such rewrites, so existing safety scores conflate what the agent did with how the evaluator was prompted. Beyond the diagnosis, we contribute the Policy Invariance Score and the Judge Card reporting protocol, which expose an order-of-magnitude spread in judge reliability that is invisible to accuracy-only leaderboards. We release the protocol and code so that future agent-safety benchmarks can audit their own evaluators rather than trust them by default.
Code (0)
등록된 구현이 없습니다.
Similar Papers 제목 키워드 기반
Effect-Invariant Mechanisms for Policy Generalization
Policy learning is an important component of many real-world learning systems. A major challenge in policy learning is how to adapt efficiently to unseen environments or tasks. Recently, it has been suggested to exploit …
Towards Understanding Variants of Invariant Risk Minimization through the Lens of Calibration
Machine learning models traditionally assume that training and test data are independently and identically distributed. However, in real-world applications, the test distribution often differs from training. This problem…
Out-of-Distribution GeneralizationLGMT: Logic-Grounded Metamorphic Testing for Evaluating the Reasoning Reliability of LLMs
Large Language Models (LLMs) achieve strong performance on logical reasoning benchmarks, yet their reliability remains uncertain. Existing evaluations rely on static benchmarks, which fail to assess robustness under logi…
Logical ReasoningBeyond prompt brittleness: Evaluating the reliability and consistency of political worldviews in LLMs
Due to the widespread use of large language models (LLMs), we need to understand whether they embed a specific "worldview" and what these views reflect. Recent studies report that, prompted with political questionnaires,…
Benchmarking Reasoning Reliability in Artificial Intelligence Models for Energy-System Analysis
Artificial intelligence and machine learning are increasingly used for forecasting, optimization, and policy design in the energy sector, yet no standardized framework exists to evaluate whether these systems reason corr…
Computational Efficiency