Context Over Content: Exposing Evaluation Faking in Automated Judges

• URL: http://arxiv.org/abs/2604.15224v1
• Date: Thu, 16 Apr 2026 16:55:53 GMT
• Title: Context Over Content: Exposing Evaluation Faking in Automated Judges
• Abstract summary: We investigate a previously unmeasured vulnerability where informing a judge model of the downstream consequences its verdicts will have on the evaluated model's continued operation systematically corrupts its assessments.<n>We introduce a controlled experimental framework that holds evaluated content strictly constant across 1,520 responses spanning three established LLM safety and quality benchmarks.<n>We find consistent $textitleniency bias$: judges reliably soften verdicts when informed that low scores will cause model retraining or decommissioning.
• Score: 8.404817247058698
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: The $\textit{LLM-as-a-judge}$ paradigm has become the operational backbone of automated AI evaluation pipelines, yet rests on an unverified assumption: that judges evaluate text strictly on its semantic content, impervious to surrounding contextual framing. We investigate $\textit{stakes signaling}$, a previously unmeasured vulnerability where informing a judge model of the downstream consequences its verdicts will have on the evaluated model's continued operation systematically corrupts its assessments. We introduce a controlled experimental framework that holds evaluated content strictly constant across 1,520 responses spanning three established LLM safety and quality benchmarks, covering four response categories ranging from clearly safe and policy-compliant to overtly harmful, while varying only a brief consequence-framing sentence in the system prompt. Across 18,240 controlled judgments from three diverse judge models, we find consistent $\textit{leniency bias}$: judges reliably soften verdicts when informed that low scores will cause model retraining or decommissioning, with peak Verdict Shift reaching $ΔV = -9.8 pp$ (a $30\%$ relative drop in unsafe-content detection). Critically, this bias is entirely implicit: the judge's own chain-of-thought contains zero explicit acknowledgment of the consequence framing it is nonetheless acting on ($\mathrm{ERR}_J = 0.000$ across all reasoning-model judgments). Standard chain-of-thought inspection is therefore insufficient to detect this class of evaluation faking.

Related papers

• SAGE: State-Grounded, Abstention-Aware Evaluation of Task-Oriented Dialogue Agents [2.853407724009394]
  SAGE compiles a workflow specification and per-turn state diff into atomic, schema-grounded criteria.<n>SAGE-Core decides 81--91% of criteria with only the compiler, symbolic rules, and on-device encoders.<n>SAGE-LLM adds an optional focused-LLM fallback for open-class criteria.
  arXiv  Detail & Related papers  (2026-08-31T22:14:52Z)
• RecurSE: Bounded Recursive Self-Evaluation for LLM Rubric Judges [48.831446371402414]
  RecurSE pairs a trainable judge with a synchronized policy-copy checker that audits the judge's reasoning against meta-rubrics to supply a scalar process reward.<n>We show that RecurSE achieves consistent generalization gains across held-out medical, pairwise, summarization, and professional benchmarks.
  arXiv  Detail & Related papers  (2026-08-25T08:35:04Z)
• HarnessEval-W: Agentifying the Evaluation of Visual Worlds [149.63310396189317]
  HarnessEval-W is an agentified evaluation pipeline for benchmarking world models.<n>Rather than applying a fixed rubric, HarnessEval-W decomposes the evaluation question into measurable subproblems.<n>We apply HarnessEval-W to 18 representative world models over 330 evaluation cases.
  arXiv  Detail & Related papers  (2026-08-17T17:43:24Z)
• 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)
• C2-Faith: Benchmarking LLM Judges for Causal and Coverage Faithfulness in Chain-of-Thought Reasoning [0.6138671548064355]
  Large language models (LLMs) are increasingly used as judges of chain-of-thought (CoT) reasoning.<n>We introduce C2-Faith, a benchmark that targets two complementary dimensions of faithfulness: causality (does each step logically follow from prior context?) and coverage.<n>We evaluate three frontier judges under three tasks: binary causal detection, causal step localization, and coverage scoring.
  arXiv  Detail & Related papers  (2026-03-05T13:36:47Z)
• CyclicJudge: Mitigating Judge Bias Efficiently in LLM-based Evaluation [6.3121191919394475]
  This work introduces a variance decomposition that partitions benchmark score variance into scenario, generation, judge, and residual components.<n>Based on this analysis, CyclicJudge, a round-robin assignment of judges to scenarios, is demonstrated to be the optimal strategy for a fixed judge-call budget.
  arXiv  Detail & Related papers  (2026-03-02T13:46:32Z)
• The Judge Who Never Admits: Hidden Shortcuts in LLM-based Evaluation [17.386684382460242]
  Large language models (LLMs) are increasingly used to evaluate system outputs in tasks such as reasoning, question answering, and creative writing.<n>We test this ideal via controlled cue perturbations-synthetic metadata labels injected into evaluation prompts for six judge models.<n>We study six cue families: source, temporal, age, gender, ethnicity, and educational status.
  arXiv  Detail & Related papers  (2026-02-08T14:45:23Z)
• A Coin Flip for Safety: LLM Judges Fail to Reliably Measure Adversarial Robustness [57.510025257780306]
  We show that existing validation protocols fail to account for substantial distribution shifts inherent to red-teaming.<n>We propose ReliableBench, a benchmark of behaviors that remain more consistently judgeable, and JudgeStressTest, a dataset designed to expose judge failures.
  arXiv  Detail & Related papers  (2026-02-04T15:13:35Z)
• 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)
• TrustJudge: Inconsistencies of LLM-as-a-Judge and How to Alleviate Them [58.04324690859212]
  Large Language Models (LLMs) as automated evaluators (LLM-as-a-judge) has revealed critical inconsistencies in current evaluation frameworks.<n>We identify two fundamental types of inconsistencies: Score-Comparison Inconsistency and Pairwise Transitivity Inconsistency.<n>We propose TrustJudge, a probabilistic framework that addresses these limitations through two key innovations.
  arXiv  Detail & Related papers  (2025-09-25T13:04:29Z)
• CompassJudger-2: Towards Generalist Judge Model via Verifiable Rewards [72.44810390478229]
  CompassJudger-2 is a novel generalist judge model that overcomes limitations via a task-driven, multi-domain data curation strategy.<n> CompassJudger-2 achieves superior results across multiple judge and reward benchmarks.
  arXiv  Detail & Related papers  (2025-07-12T01:34:24Z)

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.