mlsanityData to 14 Oct 2026

Explaining Length Bias in LLM-Based Preference Evaluations

arXiv 2407.01085 ↗First posted 1 Jul 2024cs.LG cs.CLcited by 49

Not on the map. Why?

Abstract (arXiv)

The use of large language models (LLMs) as judges, particularly in preference comparisons, has become widespread, but this reveals a notable bias towards longer responses, undermining the reliability of such evaluations. To better understand such bias, we propose to decompose the preference evaluation metric, specifically the win rate, into two key components: desirability and information mass, where the former is length-independent and related to trustworthiness such as correctness, toxicity, and consistency, and the latter is length-dependent and represents the amount of information in the response. We empirically demonstrated the decomposition through controlled experiments and found that response length impacts evaluations by influencing information mass. To derive a reliable evaluation metric that assesses content quality without being confounded by response length, we propose AdapAlpaca, a simple yet effective adjustment to win rate measurement. Specifically, AdapAlpaca ensures a fair comparison of response quality by aligning the lengths of reference and test model responses under equivalent length intervals.

Key facts. Cited by 49 papers in the map (data through 14 Oct 2026). Builds on 22 of its 52 references that are in the map. First posted on arXiv 1 Jul 2024.

Builds on

22 of 52 references.

References · 22

  1. 2022Training language models to follow instructions with human feedbackcited by 15,971
  2. 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 8,065
  3. 2022Self-Instruct: Aligning Language Models with Self-Generated Instructionscited by 2,724
  4. 2023Large Language Models are not Fair Evaluatorscited by 1,085
  5. 2023AlpacaFarm: A Simulation Framework for Methods that Learn from Human Feedbackcited by 748
  6. 2023From Quantity to Quality: Boosting LLM Performance with Self-Guided Data Selection for Instruction Tuningcited by 307
  7. 2020The Pushshift Reddit Datasetcited by 292
  8. 2023Recommender AI Agent: Integrating Large Language Models for Interactive Recommendationscited by 131
  9. 2023MoDS: Model-oriented Data Selection for Instruction Tuningcited by 107
  10. 2024Long Is More for Alignment: A Simple but Tough-to-Beat Baseline for Instruction Fine-Tuningcited by 85
  11. 2023Loose lips sink ships: Mitigating Length Bias in Reinforcement Learning from Human Feedbackcited by 78
  12. 2024Offline Training of Language Model Agents with Functions as Learnable Weightscited by 48
10 more
  1. 2024Adaptive In-conversation Team Building for Language Model Agentscited by 39
  2. 2022Spurious Correlations in Reference-Free Evaluation of Text Generationcited by 26
  3. 2020DQI: Measuring Data Quality in NLPcited by 24
  4. 2024RecAI: Leveraging Large Language Models for Next-Generation Recommender Systemscited by 20
  5. 2023Koala: An Index for Quantifying Overlaps with Pre-training Corporacited by 19
  6. 2024RECOST: External Knowledge Guided Data-efficient Instruction Tuningcited by 18
  7. 2023RecExplainer: Aligning Large Language Models for Explaining Recommendation Modelscited by 14
  8. 2023NLPBench: Evaluating Large Language Models on Solving NLP Problemscited by 8
  9. 2024Better Explain Transformers by Illuminating Important Informationcited by 8
  10. 2024Aligning Language Models for Versatile Text-based Item Retrievalcited by 4

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2024: 3; 2025: 18; 2026: 28.

49 papers.

Sort:

By area:

  1. 2024LLMs-as-Judges: A Comprehensive Survey on LLM-based Evaluation Methodscited by 465
  2. 2024JudgeBench: A Benchmark for Evaluating LLM-based Judgescited by 318
  3. 2024FlipAttack: Jailbreak LLMs via Flippingcited by 80
  4. 2025Efficient Inference for Large Reasoning Models: A Surveycited by 61
  5. 2025The Hallucination Tax of Reinforcement Finetuningcited by 40
  6. 2025GVPO: Group Variance Policy Optimization for Large Language Model Post-Trainingcited by 32
  7. 2025Beyond the Surface: Measuring Self-Preference in LLM Judgmentscited by 31
  8. 2025Assessing Judging Bias in Large Reasoning Models: An Empirical Studycited by 30
  9. 2026HealthBench Professional: Evaluating Large Language Models on Real Clinician Chatscited by 26
  10. 2025Adversarial Paraphrasing: A Universal Attack for Humanizing AI-Generated Textcited by 17
  11. 2026ArenaRL: Scaling RL for Open-Ended Agents via Tournament-based Relative Rankingcited by 17
  12. 2025λ-GRPO: Unifying the GRPO Frameworks with Learnable Token Preferencescited by 10
37 more
  1. 2026Quantifying and Mitigating Self-Preference Bias of LLM Judgescited by 10
  2. 2025Inverse Reinforcement Learning Meets Large Language Model Post-Training: Basics, Advances, and Opportunitiescited by 7
  3. 2025Think Twice: Branch-and-Rethink Reasoning Reward Modelcited by 7
  4. 2025SSR: Socratic Self-Refine for Large Language Model Reasoningcited by 7
  5. 2025Adversarial Attacks on LLM-as-a-Judge Systems: Insights from Prompt Injectionscited by 6
  6. 2026Can LLMs Write Reliable Rubrics? A Meta-Evaluation for Experiment Reproductioncited by 6
  7. 2025Building Benchmarks from the Ground Up: Community-Centered Evaluation of LLMs in Healthcare Chatbot Settingscited by 5
  8. 2026Do Chatbot LLMs Talk Too Much? The YapBench Benchmarkcited by 5
  9. 2026EMPA: Evaluating Persona-Aligned Empathy as a Processcited by 4
  10. 2025Decision-Making with Deliberation: Meta-reviewing as a Document-grounded Dialoguecited by 3
  11. 2025LLMs Can Get "Brain Rot": A Pilot Study on Twitter/Xcited by 3
  12. 2026One Bias After Another: Mechanistic Reward Shaping and Persistent Biases in Language Reward Modelscited by 3
  13. 2026CoAct: Co-Active LLM Preference Learning with Human-AI Synergycited by 3
  14. 2026When Vision-Language Models Judge Without Seeing: Exposing Informativeness Biascited by 3
  15. 2025Beyond Task-Oriented and Chitchat Dialogues: Proactive and Transition-Aware Conversational Agentscited by 2
  16. 2026When More Words Say Less: Decoupling Length and Specificity in Image Description Evaluationcited by 2
  17. 2026Designing Reliable LLM-Assisted Rubric Scoring for Constructed Responses: Evidence from Physics Examscited by 2
  18. 2026From Uncertain Judgments to Calibrated Rankings: Conformal Elo Estimation for LLM Evaluationcited by 2
  19. 2026K-Bench: measuring model performance on real scientific agent requestscited by 2
  20. 2025Unveiling the Learning Mind of Language Models: A Cognitive Framework and Empirical Studycited by 1
  21. 2025ConCISE: A Reference-Free Conciseness Evaluation Metric for LLM-Generated Answerscited by 1
  22. 2026SpreadsheetArena: Decomposing Preference in LLM Generation of Spreadsheet Workbookscited by 1
  23. 2026QuarkMedBench: A Real-World Scenario Driven Benchmark for Evaluating Large Language Modelscited by 1
  24. 2026TiCo: Time-Controllable Spoken Dialogue Modelcited by 1
  25. 2026MLLM-as-a-Judge Exhibits Model Preference Biascited by 1
  26. 2026When the Strongest Teacher Is Not the Best Teacher: Student-Centric Answer Selectioncited by 1
  27. 2026StatefulDiscovery: Evidence-Calibrated Claim Formation in Open-Ended Scientific Discoverycited by 1
  28. 2025Optimizing for Persuasion Improves LLM Generalization: Evidence from Quality-Diversity Evolution of Debate Strategiescited by 0
  29. 2026METIS: Mentoring Engine for Thoughtful Inquiry & Solutionscited by 0
  30. 2026Blind to the Human Touch: Overlap Bias in LLM-Based Summary Evaluationcited by 0
  31. 2026GRASP: Deterministic argument ranking in interaction graphscited by 0
  32. 2026MDIA: A Multi-Agent Diagnostic Intelligence Pipeline on HealthBench Professionalcited by 0
  33. 2026Generalistic or Specific Embeddings, Which is Better? An Empirical Study on Search for Clinical Coding in Non-English Languagescited by 0
  34. 2026Language Shapes Instruction Hierarchy Compliance in Multilingual LLMscited by 0
  35. 2026For Your Eyes Only: Evaluating Coordination Between Isolated Language Model Instancescited by 0
  36. 2026One Model Is Not a Crowd: Multi-LLM and Aspect-Conditioned Diverse Comment Generationcited by 0
  37. 2026Quality Determines Direction, Length Shapes Magnitude: Length Control for Open-Ended Reinforcement Learningcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026A Consensus-Based Framework for Relative Preference Evaluation of Large Language Models
  2. 2025Think Again! The Effect of Test-Time Compute on Preferences, Opinions, and Beliefs of Large Language Models
  3. 2026Evaluating LLM-Generated Preference Distributions
  4. 2024Compare without Despair: Reliable Preference Evaluation with Generation Separability
  5. 2025Fragile Preferences: A Deep Dive Into Order Effects in Large Language Models
  6. 2025Do Language Models Think Consistently? A Study of Value Preferences Across Varying Response Lengths
  7. 2025Disentangling Length Bias In Preference Learning Via Response-Conditioned Modeling
  8. 2024Self-Preference Bias in LLM-as-a-Judge