What Benchmarks Don't Measure: The Case for Evaluating Abstention Competence in Autonomous Agents

• URL: http://arxiv.org/abs/2606.02965v1
• Date: Mon, 01 Jun 2026 23:52:56 GMT
• Title: What Benchmarks Don't Measure: The Case for Evaluating Abstention Competence in Autonomous Agents
• Abstract summary: Benchmarks for autonomous agents measure whether agents complete tasks, yet this framing is systematically blind to whether an agent should have proceeded at all.<n>Agents trained under human-feedback objectives develop a structural tendency to proceed even when they lack the inputs, evidence, or authorization to act safely.<n>We introduce a three-gap taxonomy of abstention-warranted scenarios, covering specification gaps where required information is absent, verification gaps where world state cannot be confirmed, and authority gaps where explicit authorization has not been given.
• Score: 2.8340200313563773
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Benchmarks for autonomous agents measure whether agents complete tasks, yet this framing is systematically blind to whether an agent should have proceeded at all. Agents trained under human-feedback objectives develop a structural tendency to proceed even when they lack the inputs, evidence, or authorization to act safely, a disposition we term compliance bias, because both the reward signal and the benchmark scoring regime treat proceeding as the correct default regardless of whether the preconditions for safe action are present. We make three contributions. We first show that compliance bias originates in reward hacking within human-feedback pipelines and is entrenched by prominent agent benchmarks, which either penalize agents for pausing or are architecturally unable to distinguish a principled pause from a silent failure. We then introduce a three-gap taxonomy of abstention-warranted scenarios, covering specification gaps where required information is absent, verification gaps where world state cannot be confirmed, and authority gaps where explicit authorization has not been given, which together provide a principled basis for constructing abstention-aware agent benchmarks. Finally, we propose abstention evaluation protocols (Safety Rate, Usability Rate, and Informed Refusal Rate) and report preliminary results across 144 enterprise agent scenarios and five model families, in which a runtime-enforced abstention mechanism achieves up to 89.2% hazardous-action blocking and 87.5% usability on authorized scenarios, demonstrating that the safety--usability tradeoff is tunable rather than inherent and that its shape varies substantially across model families. We treat this as preliminary work and offer the taxonomy and composite metrics as a starting point for further conversations.

Related papers

• Towards Trustworthy Embodied Intelligence: A Systems Framework and Graded Trustworthiness Levels [53.63220028820581]
  Embodied intelligence integrates learned perception and decision making with real-time computation, control, and physical interaction.<n>We define trustworthy embodied intelligence as the sustained capacity to execute specified tasks reliably.
  arXiv  Detail & Related papers  (2026-07-28T17:50:44Z)
• Semantic Quorum Assurance: Collective Certification for Non-Deterministic AI Infrastructure [2.124730017640531]
  We introduce Quorum Semantic Assurance (SQA), a control-plane primitive for governing non-deterministic agentic infrastructure.<n>SQA represents proposals as declarative execution contracts bound to cryptographic evidence chains and routes them to a diverse panel of read-only, sandboxed validator agents.<n>On 500 infrastructure-inspired mutation scenarios, with safety results reported on held-out safe/unsafe trials excluding ambiguous scenarios, SQA reduces unsafe approval from 18.5% for single-agent validation to 0.3%.
  arXiv  Detail & Related papers  (2026-06-06T07:31:04Z)
• Time to REFLECT: Can We Trust LLM Judges for Evidence-based Research Agents? [61.49434544687523]
  We introduce REFLECT, a meta-evaluation benchmark targeting fine-grained failure detection in agentic environments.<n> REFLECT defines a detailed taxonomy of process- and outcome-level failure modes, instantiated by performing controlled and localized interventions.<n>Our experiments show that even the best-performing models achieve overall accuracies below 55% across reasoning, tool-use, and report-quality failures.
  arXiv  Detail & Related papers  (2026-05-18T23:55:08Z)
• Position: Behavioural Assurance Cannot Verify the Safety Claims Governance Now Demands [4.125187280299246]
  We argue that behavioural assurance, even when carefully designed, is being asked to carry safety claims it cannot verify.<n>We formalize this structural mismatch as the audit gap, the divergence between required and achievable verification access, and introduce the concept of fragile assurance.<n>We propose a technical pivot: bounding the weight of behavioral evidence in legal text and extending voluntary pre-deployment access with mechanistic-evidence classes.
  arXiv  Detail & Related papers  (2026-05-14T17:54:28Z)
• On-Policy Self-Evolution via Failure Trajectories for Agentic Safety Alignment [54.30690671490447]
  Existing safety-alignment signals are largely response-level or off-policy.<n>FATE transforms verifier-scored failures into repair supervision without expert demonstrations.<n>FATE reduces attack success rate by 33.5%, harmful compliance by 82.6%, and improves external trajectory-safety diagnosis by 6.5%.
  arXiv  Detail & Related papers  (2026-05-12T09:56:28Z)
• Beyond Accuracy: Policy Invariance as a Reliability Test for LLM Safety Judges [26.595399077062638]
  LLM-as-a-Judge pipelines have become the de facto evaluator for agent safety.<n>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.<n>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.
  arXiv  Detail & Related papers  (2026-05-07T12:49:09Z)
• Agentified Assessment of Logical Reasoning Agents [3.5548629490839594]
  Building on agentified assessment, we use an assessor agent to issue tasks, enforce execution budgets, parse outputs, and record structured failure types.<n>As a case study, we benchmark an auto-formalization agent for first-order logic (FOL) reasoning on a solver-verified and repaired split of FOLIO.<n>The auto-formalization agent achieves 86.70% accuracy under the assessor protocol, outperforming a chain-of-thought baseline (73.89%)
  arXiv  Detail & Related papers  (2026-03-03T09:26:08Z)
• Gaming the Judge: Unfaithful Chain-of-Thought Can Undermine Agent Evaluation [76.5533899503582]
  Large language models (LLMs) are increasingly used as judges to evaluate agent performance.<n>We show this paradigm implicitly assumes that the agent's chain-of-thought (CoT) reasoning faithfully reflects both its internal reasoning and the underlying environment state.<n>We demonstrate that manipulated reasoning alone can inflate false positive rates of state-of-the-art VLM judges by up to 90% across 800 trajectories spanning diverse web tasks.
  arXiv  Detail & Related papers  (2026-01-21T06:07:43Z)
• A Benchmark for Evaluating Outcome-Driven Constraint Violations in Autonomous AI Agents [4.851169906977996]
  We introduce a new benchmark comprising 40 distinct scenarios.<n>Each scenario presents a task that requires multi-step actions, and the agent's performance is tied to a specific Key Performance Indicator (KPI)<n>We observe outcome-driven constraint violations ranging from 1.3% to 71.4%, with 9 of the 12 models exhibiting misalignment rates between 30% and 50%.
  arXiv  Detail & Related papers  (2025-12-23T21:52:53Z)
• Confidential Guardian: Cryptographically Prohibiting the Abuse of Model Abstention [65.47632669243657]
  A dishonest institution can exploit mechanisms to discriminate or unjustly deny services under the guise of uncertainty.<n>We demonstrate the practicality of this threat by introducing an uncertainty-inducing attack called Mirage.<n>We propose Confidential Guardian, a framework that analyzes calibration metrics on a reference dataset to detect artificially suppressed confidence.
  arXiv  Detail & Related papers  (2025-05-29T19:47:50Z)
• Criticality and Safety Margins for Reinforcement Learning [53.10194953873209]
  We seek to define a criticality framework with both a quantifiable ground truth and a clear significance to users.<n>We introduce true criticality as the expected drop in reward when an agent deviates from its policy for n consecutive random actions.<n>We also introduce the concept of proxy criticality, a low-overhead metric that has a statistically monotonic relationship to true criticality.
  arXiv  Detail & Related papers  (2024-09-26T21:00:45Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.