mlsanityData to 7 Oct 2026

How do Large Language Models Navigate Conflicts between Honesty and Helpfulness?

arXiv 2402.07282 ↗First posted 11 Feb 2024cs.CL cs.AI cs.LGcited by 27

Not on the map. Why?

Abstract (arXiv)

In day-to-day communication, people often approximate the truth - for example, rounding the time or omitting details - in order to be maximally helpful to the listener. How do large language models (LLMs) handle such nuanced trade-offs? To address this question, we use psychological models and experiments designed to characterize human behavior to analyze LLMs. We test a range of LLMs and explore how optimization for human preferences or inference-time reasoning affects these trade-offs. We find that reinforcement learning from human feedback improves both honesty and helpfulness, while chain-of-thought prompting skews LLMs towards helpfulness over honesty. Finally, GPT-4 Turbo demonstrates human-like response patterns including sensitivity to the conversational framing and listener's decision context. Our findings reveal the conversational values internalized by LLMs and suggest that even these abstract values can, to a degree, be steered by zero-shot prompting.

Key facts. Cited by 27 papers in the map (data through 7 Oct 2026). Builds on 31 of its 62 references that are in the map. First posted on arXiv 11 Feb 2024.

Builds on

31 of 62 references.

References · 31

  1. 2022Training language models to follow instructions with human feedbackcited by 15,712
  2. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,302
  3. 2022Large Language Models are Zero-Shot Reasonerscited by 4,323
  4. 2021On the Opportunities and Risks of Foundation Modelscited by 3,961
  5. 2017Deep reinforcement learning from human preferencescited by 3,846
  6. 2022Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedbackcited by 3,634
  7. 2022Constitutional AI: Harmlessness from AI Feedbackcited by 2,950
  8. 2021TruthfulQA: Measuring How Models Mimic Human Falsehoodscited by 2,919
  9. 2020REALM: Retrieval-Augmented Language Model Pre-Trainingcited by 2,126
  10. 2022Survey of Hallucination in Natural Language Generationcited by 2,056
  11. 2022Beyond the Imitation Game: Quantifying and extrapolating the capabilities of language modelscited by 1,970
  12. 2021WebGPT: Browser-assisted question-answering with human feedbackcited by 1,722
19 more
  1. 2020RealToxicityPrompts: Evaluating Neural Toxic Degeneration in Language Modelscited by 1,235
  2. 2021A General Language Assistant as a Laboratory for Alignmentcited by 1,023
  3. 2022Improving alignment of dialogue agents via targeted human judgementscited by 527
  4. 2023Towards Measuring the Representation of Subjective Global Opinions in Language Modelscited by 424
  5. 2022Using cognitive psychology to understand GPT-3cited by 389
  6. 2018Speaker-Follower Models for Vision-and-Language Navigationcited by 337
  7. 2022Teaching language models to support answers with verified quotescited by 290
  8. 2021Internet-Augmented Dialogue Generationcited by 217
  9. 2022Talking About Large Language Modelscited by 174
  10. 2022Language models show human-like content effects on reasoning taskscited by 171
  11. 2023The Capacity for Moral Self-Correction in Large Language Modelscited by 171
  12. 2023ReviewerGPT? An Exploratory Study on Using Large Language Models for Paper Reviewingcited by 118
  13. 2023Evaluating and Mitigating Discrimination in Language Model Decisionscited by 117
  14. 2023Why think step by step? Reasoning emerges from the locality of experiencecited by 103
  15. 2022In conversation with Artificial Intelligence: aligning language models with human valuescited by 61
  16. 2024PUB: A Pragmatics Understanding Benchmark for Assessing LLMs' Pragmatics Capabilitiescited by 58
  17. 2022Can language models handle recursively nested grammatical structures? A case study on comparing models and humanscited by 41
  18. 2023Evaluating statistical language models as pragmatic reasonerscited by 14
  19. 2022Pragmatics in Language Grounding: Phenomena, Tasks, and Modeling Approachescited by 12

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 8; 2025: 15; 2026: 4.

27 papers.

Sort:

By area:

  1. 2024Mind Your Step (by Step): Chain-of-Thought can Reduce Performance on Tasks where Thinking Makes Humans Worsecited by 107
  2. 2025LLM Social Simulations Are a Promising Research Methodcited by 97
  3. 2024Large Language Models Assume People are More Rational than We Really arecited by 49
  4. 2025Accumulating Context Changes the Beliefs of Language Modelscited by 25
  5. 2024The Impossibility of Fair LLMscited by 22
  6. 2024A Survey on the Honesty of Large Language Modelscited by 21
  7. 2025Training language models to be warm and empathetic makes them less reliable and more sycophanticcited by 19
  8. 2024HonestLLM: Toward an Honest and Helpful Large Language Modelcited by 17
  9. 2024CtrlA: Adaptive Retrieval-Augmented Generation via Inherent Controlcited by 13
  10. 2025Generative Value Conflicts Reveal LLM Prioritiescited by 13
  11. 2024AI Alignment through Reinforcement Learning from Human Feedback? Contradictions and Limitationscited by 10
  12. 2025A Framework for Studying AI Agent Behavior: Evidence from Consumer Choice Experimentscited by 10
15 more
  1. 2025EconEvals: Benchmarks and Litmus Tests for Economic Decision-Making by LLM Agentscited by 8
  2. 2025Are Large Language Models Sensitive to the Motives Behind Communication?cited by 7
  3. 2025LLM Agents Are Hypersensitive to Nudgescited by 5
  4. 2025Evaluating Language Models' Evaluations of Gamescited by 4
  5. 2025LLM Agents Display Human Biases but Exhibit Distinct Learning Patternscited by 3
  6. 2025Cognitive models can reveal interpretable value trade-offs in language modelscited by 3
  7. 2025The Narrative Continuity Test: A Conceptual Framework for Evaluating Identity Persistence in AI Systemscited by 3
  8. 2026Ads in AI Chatbots? An Analysis of How Large Language Models Navigate Conflicts of Interestcited by 3
  9. 2024Wait, but Tylenol is Acetaminophen... Investigating and Improving Language Models' Ability to Resist Requests for Misinformationcited by 2
  10. 2025Accelerating RLHF Training with Reward Variance Increasecited by 2
  11. 2025PICACO: Pluralistic In-Context Value Alignment of LLMs via Total Correlation Optimizationcited by 2
  12. 2025Prompt Optimization Across Multiple Agents for Representing Diverse Human Populationscited by 1
  13. 2026Normative Robustness as a Frontier for Non-Verifiable Reasoning in LLMscited by 0
  14. 2026CogGym: Towards Large-Scale Comparative Evaluation of Human and Machine Cognitioncited by 0
  15. 2026Are Human-Aligned Models Models of Humans? A Turing-Test Gap in Preference Alignmentcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Human Preferences for Constructive Interactions in Language Model Alignment
  2. 2024Dishonesty in Helpful and Harmless Alignment
  3. 2024Comparing Rationality Between Large Language Models and Humans: Insights and Open Questions
  4. 2023Alignment for Honesty
  5. 2024AI-LieDar: Examine the Trade-off Between Utility and Truthfulness in LLM Agents
  6. 2026Think Before You Lie: How Reasoning Leads to Honesty
  7. 2026The Fragility Of Moral Judgment In Large Language Models
  8. 2025Alignment Revisited: Are Large Language Models Consistent in Stated and Revealed Preferences?