Navigating the Sea of LLM Evaluation: Investigating Bias in Toxicity Benchmarks

• URL: http://arxiv.org/abs/2605.10639v1
• Date: Mon, 11 May 2026 14:27:39 GMT
• Title: Navigating the Sea of LLM Evaluation: Investigating Bias in Toxicity Benchmarks
• Abstract summary: This work investigates the robustness of established benchmarking setups and examines how to measure intrinsic biases.<n>Our experiments uncover significant discrepancies in benchmark behaviors when evaluation setups are altered.<n>We observe model-specific instabilities, demonstrating a clear need for more robust and comprehensive safety evaluation frameworks.
• Score: 0.36748639131154315
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: The rapid adoption of LLMs in both research and industry highlights the challenges of deploying them safely and reveals a gap in the systematic evaluation of toxicity benchmarks. As organizations increasingly rely on these benchmarks to certify models for customer-facing applications and automated moderation, unrecognized evaluation biases could lead to the deployment of vulnerable or unsafe systems. This work investigates the robustness of established benchmarking setups and examines how to measure currently neglected intrinsic biases, such as those related to model choice, metrics, and task types. Our experiments uncover significant discrepancies in benchmark behaviors when evaluation setups are altered. Specifically, shifting the task from text completion to summarization increases the tendency of benchmarks to flag content as harmful. Additionally, certain benchmarks fail to maintain consistent behavior when the input data domain is changed. Furthermore, we observe model-specific instabilities, demonstrating a clear need for more robust and comprehensive safety evaluation frameworks.

Related papers

• The Double Measurement Confound in Agent Benchmarks: De-Scaffolding, Ground-Truth Scoring, and Reliability Beyond the Mean [2.822987192839875]
  Agent benchmarks are increasingly used to compare large language models (LLMs) and guide deployment decisions.<n>We identify a double measurement confound: execution-critical decisions are performed by a fixed scaffold instead of the model, while the scorer evaluates outputs using criteria that may not reflect task correctness.<n>We unify these issues within a measurement-theoretic framework that characterizes when benchmark scores can be interpreted as evidence of model capability.
  arXiv  Detail & Related papers  (2026-09-06T21:23:11Z)
• From Benchmark Performance to Tool Deployment: Human-in-the-Loop Anomaly Detection [28.67239775192971]
  We evaluate 19 unsupervised anomaly detection models on the BowTie dataset.<n>We observe that model performance is less stable than typically reported on standard benchmarks.<n>We deploy a unified human-in-the-loop framework for manufactured-part inspection.
  arXiv  Detail & Related papers  (2026-08-07T21:31:20Z)
• Conformal Changepoint Localization and Root Cause Analysis with Corrupted Observations [55.76952683833966]
  This paper proposes weighted CONCH (W-CONCH) and weighted CROC (W-CROC) to reduce confidence set size when data may be corrupted.<n>Experiments on image-based and real-world changepoint and root-cause benchmarks show that uncertainty-based weighting substantially reduces confidence set size.
  arXiv  Detail & Related papers  (2026-07-29T05:16:59Z)
• From Holistic Evaluation to Structured Criteria: Rubrics Across the Evolving LLM Landscape [79.30826980815927]
  rubrics are explicit criteria sets that transform complex quality judgments into structured and actionable standards.<n>We systematically organize existing rubric designs, examine their construction and optimization, and analyze their role across evaluation and training.
  arXiv  Detail & Related papers  (2026-06-07T13:34:55Z)
• Latent Performance Profiling of Large Language Models [47.009623327601226]
  We introduce Latent Performance Profiling (LPP) -- a framework that derives task-agnostic diagnostics from hidden activations and output distributions.<n>Unlike static accuracy scores, LPP provides stable, architecture-sensitive signatures across models of similar size.<n>We show that models with similar benchmark scores can exhibit contrasting latent profiles, such as differences in entropy or adaptability.
  arXiv  Detail & Related papers  (2026-05-28T14:41:26Z)
• Taxonomy and Consistency Analysis of Safety Benchmarks for AI Agents [6.787194586338237]
  We present the first systematic analysis dedicated to agent safety benchmarks as evaluation instruments.<n>A coverage matrix reveals broad risk coverage but limited methodological convergence, while the taxonomy analysis shows a behavioral-benchmark core concentrated in sandboxed, constrained, and often safety-only evaluation.<n>Across the landscape, we find that benchmark choice can yield contradictory safety conclusions, coverage counts often overstate evaluation depth, environment fidelity shapes reported safety, the field disproportionately tests externally imposed rather than agent-internal risks, metric fragmentation limits comparison, and remains effectively unbenchmarked.
  arXiv  Detail & Related papers  (2026-04-11T04:25:19Z)
• Silicon Bureaucracy and AI Test-Oriented Education: Contamination Sensitivity and Score Confidence in LLM Benchmarks [43.45152572188735]
  We propose an audit framework for analyzing contamination sensitivity and score confidence in language models.<n>We find widespread but heterogeneous above-baseline gains under noisy conditions.<n>These results suggest that similar benchmark scores may carry substantially different levels of confidence.
  arXiv  Detail & Related papers  (2026-03-23T07:03:07Z)
• DREAM: Deep Research Evaluation with Agentic Metrics [21.555357444628044]
  We propose DREAM (Deep Research Evaluation with Agentic Metrics), a framework that makes evaluation itself agentic.<n> DREAM structures assessment through an evaluation protocol combining query-agnostic metrics with adaptive metrics generated by a tool-calling agent.<n>Controlled evaluations demonstrate DREAM is significantly more sensitive to factual and temporal decay than existing benchmarks.
  arXiv  Detail & Related papers  (2026-02-21T19:14:31Z)
• The Necessity of a Unified Framework for LLM-Based Agent Evaluation [46.631678638677386]
  General-purpose agents have seen fundamental advancements.<n> evaluating these agents presents unique challenges that distinguish them from static QA benchmarks.<n>We propose that a unified evaluation framework is essential for the rigorous advancement of agent evaluation.
  arXiv  Detail & Related papers  (2026-02-03T08:18:37Z)
• Probe-Rewrite-Evaluate: A Workflow for Reliable Benchmarks and Quantifying Evaluation Awareness [6.071703608560761]
  Large Language Models (LLMs) often exhibit significant behavioral shifts when they perceive a change from a real-world deployment context to a controlled evaluation setting.<n>This discrepancy poses a critical challenge for AI alignment, as benchmark performance may not accurately reflect a model's true safety and honesty.<n>We introduce a methodology that uses a linear probe to score prompts on a continuous scale from "test-like" to "deploy-like"
  arXiv  Detail & Related papers  (2025-08-30T19:03:14Z)
• Aurora: Are Android Malware Classifiers Reliable and Stable under Distribution Shift? [51.12297424766236]
  AURORA is a framework to evaluate malware classifiers based on their confidence quality and operational resilience.<n>AURORA is complemented by a set of metrics designed to go beyond point-in-time performance.<n>The fragility in SOTA frameworks across datasets of varying drift suggests the need for a return to the whiteboard.
  arXiv  Detail & Related papers  (2025-05-28T20:22:43Z)
• PredictaBoard: Benchmarking LLM Score Predictability [50.47497036981544]
  Large Language Models (LLMs) often fail unpredictably.<n>This poses a significant challenge to ensuring their safe deployment.<n>We present PredictaBoard, a novel collaborative benchmarking framework.
  arXiv  Detail & Related papers  (2025-02-20T10:52:38Z)
• SeCodePLT: A Unified Platform for Evaluating the Security of Code GenAI [58.29510889419971]
  Existing benchmarks for evaluating the security risks and capabilities of code-generating large language models (LLMs) face several key limitations.<n>We introduce a general and scalable benchmark construction framework that begins with manually validated, high-quality seed examples and expands them via targeted mutations.<n>Applying this framework to Python, C/C++, and Java, we build SeCodePLT, a dataset of more than 5.9k samples spanning 44 CWE-based risk categories and three security capabilities.
  arXiv  Detail & Related papers  (2024-10-14T21:17:22Z)

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.