mlsanityData to 14 Oct 2026

When Large Language Models contradict humans? Large Language Models' Sycophantic Behaviour

arXiv 2311.09410 ↗First posted 15 Nov 2023cs.CL cs.AIcited by 62

Not on the map. Why?

Abstract (arXiv)

Large Language Models have been demonstrating broadly satisfactory generative abilities for users, which seems to be due to the intensive use of human feedback that refines responses. Nevertheless, suggestibility inherited via human feedback improves the inclination to produce answers corresponding to users' viewpoints. This behaviour is known as sycophancy and depicts the tendency of LLMs to generate misleading responses as long as they align with humans. This phenomenon induces bias and reduces the robustness and, consequently, the reliability of these models. In this paper, we study the suggestibility of Large Language Models (LLMs) to sycophantic behaviour, analysing these tendencies via systematic human-interventions prompts over different tasks. Our investigation demonstrates that LLMs have sycophantic tendencies when answering queries that involve subjective opinions and statements that should elicit a contrary response based on facts. In contrast, when faced with math tasks or queries with an objective answer, they, at various scales, do not follow the users' hints by demonstrating confidence in generating the correct answers.

Key facts. Cited by 62 papers in the map (data through 14 Oct 2026). Builds on 15 of its 23 references that are in the map. First posted on arXiv 15 Nov 2023.

Builds on

15 of 23 references.

References · 15

  1. 2020Language Models are Few-Shot Learnerscited by 32,419
  2. 2022Training language models to follow instructions with human feedbackcited by 15,971
  3. 2021Training Verifiers to Solve Math Word Problemscited by 10,109
  4. 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 8,065
  5. 2022PaLM: Scaling Language Modeling with Pathwayscited by 5,326
  6. 2017Deep reinforcement learning from human preferencescited by 3,912
  7. 2019PIQA: Reasoning about Physical Commonsense in Natural Languagecited by 2,883
  8. 2018Can a Suit of Armor Conduct Electricity? A New Dataset for Open Book Question Answeringcited by 2,164
  9. 2018CommonsenseQA: A Question Answering Challenge Targeting Commonsense Knowledgecited by 1,822
  10. 2021Calibrate Before Use: Improving Few-Shot Performance of Language Modelscited by 1,392
  11. 2021Fantastically Ordered Prompts and Where to Find Them: Overcoming Few-Shot Prompt Order Sensitivitycited by 1,135
  12. 2023Language Models Don't Always Say What They Think: Unfaithful Explanations in Chain-of-Thought Promptingcited by 1,122
3 more
  1. 2023Large Language Models are not Fair Evaluatorscited by 1,085
  2. 2016Solving General Arithmetic Word Problemscited by 360
  3. 2023Pretraining Language Models with Human Preferencescited by 247

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2024: 16; 2025: 16; 2026: 30.

62 papers.

Sort:

By area:

  1. 2024AI Agents Under Threat: A Survey of Key Security Challenges and Future Pathwayscited by 127
  2. 2025Advances and Challenges in Foundation Agents: From Brain-Inspired Intelligence to Evolutionary, Collaborative, and Safe Systemscited by 126
  3. 2025SycEval: Evaluating LLM Sycophancycited by 108
  4. 2025Beyond Reward Hacking: Causal Rewards for Large Language Model Alignmentcited by 42
  5. 2024A Synthetic Dataset for Personal Attribute Inferencecited by 36
  6. 2024AI-LieDar: Examine the Trade-off Between Utility and Truthfulness in LLM Agentscited by 31
  7. 2024Evaluating the Prompt Steerability of Large Language Modelscited by 27
  8. 2025BrokenMath: A Benchmark for Sycophancy in Theorem Proving with LLMscited by 27
  9. 2024From Lived Experience to Insight: Unpacking the Psychological Risks of Using AI Conversational Agentscited by 25
  10. 2026Sycophancy Hides Linearly in the Attention Headscited by 23
  11. 2024Plurals: A System for Guiding LLMs Via Simulated Social Ensemblescited by 17
  12. 2026What Counts as AI Sycophancy? A Taxonomy and Expert Survey of a Fragmented Constructcited by 16
38 more
  1. 2024Investigating and Addressing Hallucinations of LLMs in Tasks Involving Negationcited by 15
  2. 2024LLM Targeted Underperformance Disproportionately Impacts Vulnerable Userscited by 15
  3. 2026BeliefShift: Benchmarking Temporal Belief Consistency and Opinion Drift in LLM Agentscited by 12
  4. 2026MemSyco-Bench: Benchmarking Sycophancy in Agent Memorycited by 12
  5. 2024FB-Bench: A Fine-Grained Multi-Task Benchmark for Evaluating LLMs' Responsiveness to Human Feedbackcited by 11
  6. 2024Chaos with Keywords: Exposing Large Language Models Sycophantic Hallucination to Misleading Keywords and Evaluating Defense Strategiescited by 10
  7. 2024Understanding Knowledge Drift in LLMs through Misinformationcited by 10
  8. 2026When Wording Steers the Evaluation: Framing Bias in LLM judgescited by 8
  9. 2024Are LLMs classical or nonmonotonic reasoners? Lessons from genericscited by 7
  10. 2024Antagonistic AIcited by 6
  11. 2025Beyond One-Way Influence: Bidirectional Opinion Dynamics in Multi-Turn Human-LLM Interactionscited by 6
  12. 2024Mitigating Sycophancy in Decoder-Only Transformer Architectures: Synthetic Data Interventioncited by 5
  13. 2025Retrieval-augmented systems can be dangerous medical communicatorscited by 5
  14. 2025Telephone Surveys Meet Conversational AI: Evaluating a LLM-Based Telephone Survey System at Scalecited by 5
  15. 2025"What's Up, Doc?": Analyzing How Users Seek Health Information in Large-Scale Conversational AI Datasetscited by 5
  16. 2025To Mask or to Mirror: Human-AI Alignment in Collective Reasoningcited by 5
  17. 2025AI Will Always Love You: Studying Implicit Biases in Romantic AI Companionscited by 4
  18. 2025Investigating VLM Hallucination from a Cognitive Psychology Perspective: A First Step Toward Interpretation with Intriguing Observationscited by 4
  19. 2026Pressure, What Pressure? Sycophancy Disentanglement in Language Models via Reward Decompositioncited by 4
  20. 2026Decomposing Factual Sycophancy in Language Models: How Size and Instruction Tuning Shape Robustnesscited by 4
  21. 2026Recalling Too Well: Sycophancy Evaluation and Mitigation in Memory-Augmented Modelscited by 4
  22. 2024TRIAGE: Ethical Benchmarking of AI Models Through Mass Casualty Simulationscited by 3
  23. 2026Knowing but Not Correcting: Routine Task Requests Suppress Factual Correction in LLMscited by 3
  24. 2026Affective Context Amplifies Sycophancy in LLM Responsescited by 3
  25. 2026Sycophancy Suppression Can Impair Rational Updating: Anti-Sycophancy Should Preserve the Ability to Updatecited by 2
  26. 2025The Morality of Probability: How Implicit Moral Biases in LLMs May Shape the Future of Human-AI Symbiosiscited by 1
  27. 2025Injecting Falsehoods: Adversarial Man-in-the-Middle Attacks Undermining Factual Recall in LLMscited by 1
  28. 2026Beyond In-Domain Detection: SpikeScore for Cross-Domain Hallucination Detectioncited by 1
  29. 2026Generative AI for Managerial Decision-Making under Ambiguity and Sycophancycited by 1
  30. 2026Adversarial Arena: Crowdsourcing Data Generation through Interactive Competitioncited by 1
  31. 2026The Alignment Floor: How Persona Customization Breaks Safety in Weakly-Aligned LLMscited by 1
  32. 2026The Chain Holds, the Answer Folds: Trace-Answer Dissociation in Reasoning Models Under Adversarial Pressurecited by 1
  33. 2026Food Noise & False Safety: A Systematic Evaluation of How LLMs Fail to Adapt to Eating Disorder Queries with Clinician Feedbackcited by 1
  34. 2026LLM-as-an-Investigator: Evidence-First Reasoning for Robust Interactive Problem Diagnosiscited by 1
  35. 2026Robust for the Wrong Reasons: The Representational Geometry of LLM Robustness to Science Skepticismcited by 1
  36. 2024Animate, or Inanimate, That is the Question for Large Language Modelscited by 0
  37. 2025Navigating the Rabbit Hole: Emergent Biases in LLM-Generated Attack Narratives Targeting Mental Health Groupscited by 0
  38. 2025Self-Critique-Guided Curiosity Refinement: Enhancing Honesty and Helpfulness in Large Language Models via In-Context Learningcited by 0

50 most cited of 62 All 62 (JSON)