Pinned and Still Unstable: Within-Judge Verdict Variance and the Noise Floor of LLM-as-Judge Leaderboards

arXiv:2609.33044 · cs.CL · Submitted 2026-09-27 · Read on arXiv

cs.CL

Submitted: 2026-09-27

Updated: 2026-09-30

Terminology

• Benchmark — 62× in this paper · explained in 2 episodes
• Consistency — 3× in this paper · explained in 3 episodes
• Generalization — 3× in this paper · explained in 13 episodes
• Provenance — 3× in this paper · explained in 3 episodes
• Robustness — 3× in this paper · explained in 5 episodes
• Alignment — 2× in this paper · explained in 1 episode
• LLM judge — 2× in this paper · explained in 2 episodes
• LLM-as-a-Judge — 2× in this paper · explained in 3 episodes
• Accuracy — 1× in this paper · explained in 1 episode
• Adversarial Vulnerabilities — 1× in this paper · explained in 1 episode
• Aggregation — 1× in this paper · explained in 1 episode
• Alignment Techniques — 1× in this paper · explained in 1 episode
• Clustering — 1× in this paper · explained in 2 episodes
• Model-Agnostic — 1× in this paper · explained in 2 episodes
• Perplexity — 1× in this paper · explained in 3 episodes
• Prompt Engineering — 1× in this paper · explained in 4 episodes
• Trace — 1× in this paper · explained in 2 episodes
• Verification — 1× in this paper · explained in 2 episodes

Sources

• Non-Determinism of "Deterministic" LLM Settings
• Position: Don't Use the CLT in LLM Evals With Fewer Than a Few Hundred Datapoints
• Humans or LLMs as the Judge? A Study on Judgement Biases
• Safer or Luckier? LLMs as Safety Evaluators Are Not Robust to Artifacts
• How is ChatGPT's behavior changing over time?
• Chatbot Arena: An Open Platform for Evaluating LLMs by Human Preference
• Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities
• AlpacaFarm: A Simulation Framework for Methods that Learn from Human Feedback
• Length-Controlled AlpacaEval: A Simple Way to Debias Automatic Evaluators
• LZ Penalty: An information-theoretic repetition penalty for autoregressive language models
• Do Repetitions Matter? Strengthening Reliability in LLM Evaluations
• From Crowdsourced Data to High-Quality Benchmarks: Arena-Hard and BenchBuilder Pipeline
• Aligning with Human Judgement: The Role of Pairwise Preference in Large Language Model Evaluators
• LLM Evaluators Recognize and Favor Their Own Generations
• Verbosity Bias in Preference Labeling by Large Language Models
• Can You Trust LLM Judgments? Reliability of LLM-as-a-Judge
• Large Language Models are Inconsistent and Biased Evaluators
• Let Me Speak Freely? A Study on the Impact of Format Restrictions on Performance of Large Language Models
• Judging the Judges: Evaluating Alignment and Vulnerabilities in LLMs-as-Judges
• The Verification Tax: Fundamental Limits of AI Auditing in the Rare-Error Regime

Related papers

• Exploring Solution Divergence and Its Effect on Large Language Model Problem Solving
• Ishigaki-IDS-Bench: A Benchmark for Generating Information Delivery Specification from BIM Information Requirements
• Subliminal Steering: Stronger Encoding of Hidden Signals
• MedStruct-S: A Benchmark for Key Discovery, Key-Conditioned QA and Semi-Structured Extraction from OCR Clinical Reports
• The End of Transformers? On Challenging Attention and the Rise of Sub-Quadratic Architectures
• Untangling the Mechanisms of Misleading Context in Medical Question Answering