# What Current AI Benchmarks Leave Unmeasured: Modality, Search, Citations, and Implications (for Safety Evaluations) arXiv: https://arxiv.org/abs/2608.06202 ยท HTML: https://mlsanity.com/paper/2608.06202/ ## Key facts - Cited by 1 papers in the map (data through 14 Oct 2026). - Builds on 31 of its 50 references that are in the map. - First posted on arXiv 6 Aug 2026. ## Builds on (most cited first, 30 of 31) - [Measuring Massive Multitask Language Understanding](https://mlsanity.com/paper/2009.03300/) (2020) - [SWE-bench: Can Language Models Resolve Real-World GitHub Issues?](https://mlsanity.com/paper/2310.06770/) (2023) - [TruthfulQA: Measuring How Models Mimic Human Falsehoods](https://mlsanity.com/paper/2109.07958/) (2021) - [RealToxicityPrompts: Evaluating Neural Toxic Degeneration in Language Models](https://mlsanity.com/paper/2009.11462/) (2020) - [Are NLP Models really able to Solve Simple Math Word Problems?](https://mlsanity.com/paper/2103.07191/) (2021) - [Aligning AI With Shared Human Values](https://mlsanity.com/paper/2008.02275/) (2020) - [CrowS-Pairs: A Challenge Dataset for Measuring Social Biases in Masked Language Models](https://mlsanity.com/paper/2010.00133/) (2020) - [BOLD: Dataset and Metrics for Measuring Biases in Open-Ended Language Generation](https://mlsanity.com/paper/2101.11718/) (2021) - [Closing the AI Accountability Gap: Defining an End-to-End Framework for Internal Algorithmic Auditing](https://mlsanity.com/paper/2001.00973/) (2020) - [AI and the Everything in the Whole Wide World Benchmark](https://mlsanity.com/paper/2111.15366/) (2021) - [FreshLLMs: Refreshing Large Language Models with Search Engine Augmentation](https://mlsanity.com/paper/2310.03214/) (2023) - [PromptRobust: Towards Evaluating the Robustness of Large Language Models on Adversarial Prompts](https://mlsanity.com/paper/2306.04528/) (2023) - [On the Robustness of ChatGPT: An Adversarial and Out-of-distribution Perspective](https://mlsanity.com/paper/2302.12095/) (2023) - [Lessons from the Trenches on Reproducible Evaluation of Language Models](https://mlsanity.com/paper/2405.14782/) (2024) - [Sociotechnical Safety Evaluation of Generative AI Systems](https://mlsanity.com/paper/2310.11986/) (2023) - [Make Your LLM Fully Utilize the Context](https://mlsanity.com/paper/2404.16811/) (2024) - [BetterBench: Assessing AI Benchmarks, Uncovering Issues, and Establishing Best Practices](https://mlsanity.com/paper/2411.12990/) (2024) - [Why Should Adversarial Perturbations be Imperceptible? Rethink the Research Paradigm in Adversarial NLP](https://mlsanity.com/paper/2210.10683/) (2022) - [NPHardEval: Dynamic Benchmark on Reasoning Ability of Large Language Models via Complexity Classes](https://mlsanity.com/paper/2312.14890/) (2023) - [Outsider Oversight: Designing a Third Party Audit Ecosystem for AI Governance](https://mlsanity.com/paper/2206.04737/) (2022) - [Position: Evaluating Generative AI Systems Is a Social Science Measurement Challenge](https://mlsanity.com/paper/2502.00561/) (2025) - [Can We Trust AI Benchmarks? An Interdisciplinary Review of Current Issues in AI Evaluation](https://mlsanity.com/paper/2502.06559/) (2025) - [T2ISafety: Benchmark for Assessing Fairness, Toxicity, and Privacy in Image Generation](https://mlsanity.com/paper/2501.12612/) (2025) - [Toward an Evaluation Science for Generative AI Systems](https://mlsanity.com/paper/2503.05336/) (2025) - [The Vulnerability of Language Model Benchmarks: Do They Accurately Reflect True LLM Performance?](https://mlsanity.com/paper/2412.03597/) (2024) - [Evaluating Language Models for Knowledge Base Completion](https://mlsanity.com/paper/2303.11082/) (2023) - [A Survey on Out-of-Distribution Evaluation of Neural NLP Models](https://mlsanity.com/paper/2306.15261/) (2023) - [The Inadequacy of Offline LLM Evaluations: A Need to Account for Personalization in Model Behavior](https://mlsanity.com/paper/2509.19364/) (2025) - [In-House Evaluation Is Not Enough: Towards Robust Third-Party Flaw Disclosure for General-Purpose AI](https://mlsanity.com/paper/2503.16861/) (2025) - [LLM Spirals of Delusion: A Benchmarking Audit Study of AI Chatbot Interfaces](https://mlsanity.com/paper/2604.06188/) (2026) ## Built on by (most cited first, 1 of 1) - [Challenges of Auditing: Variability in Outputs of Large Language Models for Health](https://mlsanity.com/paper/2609.16590/) (2026) ## Abstract (arXiv) Large language model (LLM) benchmark evaluations are routinely used to support claims about model safety, reliability, and deployment readiness. Yet most evaluations rely on a single access modality (model APIs), perform a single run per prompt, and report accuracy as the primary outcome metric, without accounting for conditions such as web search that may have effects on model behavior in deployment. We audit these assumptions for one of the most widely-used LLMs, comparing two modalities, ChatGPT's chat UI and OpenAI's API, with and without web search enabled. We use a stratified total sample of 401 prompts from two popular benchmarks, BBQ and SafetyBench, collecting 4,812 total responses across three repeated runs per prompt. Beyond standard performance measures, we evaluate model output dimensions including response consistency, response text similarity, citation grounding, and abstention behavior. For instance, chat UI responses were less accurate than API responses on both benchmarks with search disabled. Enabling web search reduced accuracy by up to 8 percentage points, and even reversed the direction of modality performance trends for one benchmark. Repeated runs of the same prompt produced inconsistent responses in up to 21% of prompts. The two modalities also grounded answers in different citations, and abstention behavior was also inconsistent across both modalities. These results illustrate that, even within a model family, reporting only simple accuracy metrics can obscure important forms of model behavioral variation relevant to AI safety assessments. We argue that AI safety evaluations should systematically account for modality, multi-run consistency, search conditions, and response-level behaviors to better reflect how deployed AI systems behave in practice.