# Hidden Measurement Error in LLM Pipelines Distorts Annotation, Evaluation, and Benchmarking

arXiv: https://arxiv.org/abs/2604.11581 · HTML: https://mlsanity.com/paper/2604.11581/

## Key facts

- Cited by 10 papers in the map (data through 14 Oct 2026).
- Builds on 42 of its 81 references that are in the map.
- First posted on arXiv 13 Apr 2026.

## Builds on (most cited first, 30 of 42)

- [Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena](https://mlsanity.com/paper/2306.05685/) (2023)
- [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)
- [Chatbot Arena: An Open Platform for Evaluating LLMs by Human Preference](https://mlsanity.com/paper/2403.04132/) (2024)
- [Length-Controlled AlpacaEval: A Simple Way to Debias Automatic Evaluators](https://mlsanity.com/paper/2404.04475/) (2024)
- [ChatGPT Outperforms Crowd-Workers for Text-Annotation Tasks](https://mlsanity.com/paper/2303.15056/) (2023)
- [Quantifying Language Models' Sensitivity to Spurious Features in Prompt Design or: How I learned to start worrying about prompt formatting](https://mlsanity.com/paper/2310.11324/) (2023)
- [From Crowdsourced Data to High-Quality Benchmarks: Arena-Hard and BenchBuilder Pipeline](https://mlsanity.com/paper/2406.11939/) (2024)
- [State of What Art? A Call for Multi-Prompt LLM Evaluation](https://mlsanity.com/paper/2401.00595/) (2023)
- [tinyBenchmarks: evaluating LLMs with fewer examples](https://mlsanity.com/paper/2402.14992/) (2024)
- [Can Large Language Models Transform Computational Social Science?](https://mlsanity.com/paper/2305.03514/) (2023)
- [The Effect of Sampling Temperature on Problem Solving in Large Language Models](https://mlsanity.com/paper/2402.05201/) (2024)
- [Adding Error Bars to Evals: A Statistical Approach to Language Model Evaluations](https://mlsanity.com/paper/2411.00640/) (2024)
- [Towards a Science of Scaling Agent Systems](https://mlsanity.com/paper/2512.08296/) (2025)
- [When Benchmarks are Targets: Revealing the Sensitivity of Large Language Model Leaderboards](https://mlsanity.com/paper/2402.01781/) (2024)
- [Benchmarking Benchmark Leakage in Large Language Models](https://mlsanity.com/paper/2404.18824/) (2024)
- [The Alternative Annotator Test for LLM-as-a-Judge: How to Statistically Justify Replacing Human Annotators with LLMs](https://mlsanity.com/paper/2501.10970/) (2025)
- [The Leaderboard Illusion](https://mlsanity.com/paper/2504.20879/) (2025)
- [Foot-In-The-Door: A Multi-turn Jailbreak for LLMs](https://mlsanity.com/paper/2502.19820/) (2025)
- [Understanding and Mitigating Numerical Sources of Nondeterminism in LLM Inference](https://mlsanity.com/paper/2506.09501/) (2025)
- [Large Language Model Hacking: Quantifying the Hidden Risks of Using LLMs for Text Annotation](https://mlsanity.com/paper/2509.08825/) (2025)
- [Cheating Automatic LLM Benchmarks: Null Models Achieve High Win Rates](https://mlsanity.com/paper/2410.07137/) (2024)
- [UProp: Investigating the Uncertainty Propagation of LLMs in Multi-Step Agentic Decision-Making](https://mlsanity.com/paper/2506.17419/) (2025)
- [Exploring the Impact of Temperature on Large Language Models:Hot or Cold?](https://mlsanity.com/paper/2506.07295/) (2025)
- [Using Imperfect Surrogates for Downstream Inference: Design-based Supervised Learning for Social Science Applications of Large Language Models](https://mlsanity.com/paper/2306.04746/) (2023)
- [Neither Valid nor Reliable? Investigating the Use of LLMs as Judges](https://mlsanity.com/paper/2508.18076/) (2025)
- [Rating Roulette: Self-Inconsistency in LLM-As-A-Judge Frameworks](https://mlsanity.com/paper/2510.27106/) (2025)
- [AILuminate: Introducing v1.0 of the AI Risk and Reliability Benchmark from MLCommons](https://mlsanity.com/paper/2503.05731/) (2025)
- [Brittlebench: Quantifying LLM robustness via prompt sensitivity](https://mlsanity.com/paper/2603.13285/) (2026)
- [Prompt Stability Scoring for Text Annotation with Large Language Models](https://mlsanity.com/paper/2407.02039/) (2024)

## Built on by (most cited first, 10 of 10)

- [TOKI: A Bitemporal Operator Algebra for Contradiction Resolution in LLM-Agent Persistent Memory](https://mlsanity.com/paper/2606.06240/) (2026)
- [Deployment Decision Reliability: A Generalizability-Theory Framework for Sizing Long-Horizon Agent Evaluations](https://mlsanity.com/paper/2608.11323/) (2026)
- [CAFE: A Compound-AI Factorial Evaluation Framework](https://mlsanity.com/paper/2607.10380/) (2026)
- [Validating LLMs in social science: Epistemic threats and emerging norms](https://mlsanity.com/paper/2607.07915/) (2026)
- [Ask Which, Not How Good: Sizing Benchmarks Scored by an LLM](https://mlsanity.com/paper/2609.27787/) (2026)
- [How Sensitive Are LLM Leaderboard Claims to Hidden Model Selection?](https://mlsanity.com/paper/2609.28177/) (2026)
- [How Reproducible Are Evaluation Conclusions? A Self-Audit of LLM-Inferred Prompt Structure](https://mlsanity.com/paper/2609.30074/) (2026)
- [What Does a Harness Repair? A Preregistered Study of Visibility, Baseline Adequacy and Evaluation Defects](https://mlsanity.com/paper/2610.05533/) (2026)
- [Truncation Samplers Are Nested: Comparing Decoding Rules Is a Measurement Problem](https://mlsanity.com/paper/2610.15570/) (2026)
- [Is the Platform Part of the Measurement? A Protocol and Simulation Study for Cross-Platform Equivalence of Machine-Labelled Policy Sentiment](https://mlsanity.com/paper/2610.16757/) (2026)

## Abstract (arXiv)

LLM evaluations drive which models get deployed, what safety standards get adopted, which research conclusions get published, and how projections of AI's labor-market impact get made. Yet standard confidence intervals ignore variability from judge model choice, model temperature, and prompt phrasing, producing under-coverage that worsens with more data. The omitted variance can shift results enough to reverse conclusions citepbaumann2025llmhacking, huang2026dropping; pipelines that fail to average over it leave the surface that ``benchmark hacking'' exploits citepsingh2025leaderboard. This paper decomposes LLM pipeline uncertainty into its sources, distinguishes variance that shrinks with more data from sensitivity to researcher design choices, and uses design-study projections to reduce total evaluation error (TEE). Across the demonstrations, naive standard errors are 40 - 60% smaller than the TEE-corrected SE. Using Chatbot Arena data, we show naive 95% CI coverage drops as n grows while TEE-corrected coverage holds at 95%, and TEE-guided pipelines restrict the benchmark gaming surface from 56 to 32 Elo (K=27), below the human-leaderboard baseline. We show further that a small pilot recovers honest CIs and projects which design changes most improve precision. Acting on those projections halves MMLU estimation error against the answer key at equivalent cost, and raises per-match agreement with human votes by 7.9 percentage points on Chatbot Arena.