mlsanityData to 7 Oct 2026

Only Say What You Know: Calibration-Aware Generation for Long-Form Factuality

arXiv 2605.01749 ↗First posted 3 May 2026cs.CLcited by 0

On the map: LLM reasoning → Test-time compute and efficient thinking

Abstract (arXiv)

Large Reasoning Models achieve strong performance on complex tasks but remain prone to hallucinations, particularly in long-form generation where errors compound across reasoning steps. Existing approaches to improving factuality, including abstention and factuality-driven optimization, follow a coupled exploration-commitment paradigm, in which intermediate reasoning is unconditionally propagated to the final output, limiting fine-grained control over information selection and integration. In this paper, we propose an Exploration-Commitment Decoupling paradigm that disentangles knowledge exploration from final commitment, enabling models to explore with awareness while answering cautiously. We instantiate the paradigm with Calibration-Aware Generation (CAG), a framework that equips models with end-to-end, calibration-aware generation capabilities, by augmenting intermediate reasoning with calibrated reliability estimates and prioritizing reliable content in final outputs. Across five long-form factuality benchmarks and multiple model families, CAG improves factuality by up to 13%, while reducing decoding time by up to 37%. Overall, our work highlights decoupling as a principled approach for more reliable long-form generation, offering directions for trustworthy and self-aware generative systems.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Test-time compute and efficient thinking, one of 17 directions in LLM reasoning. Builds on 32 of its 35 references that are in the map. First posted on arXiv 3 May 2026.

Builds on

32 of 35 references.

References · 32

  1. 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 8,851
  2. 2023Judging LLM-as-a-Judge with MT-Bench and Chatbot Arenacited by 8,393
  3. 2023GPQA: A Graduate-Level Google-Proof Q&A Benchmarkcited by 2,748
  4. 2023Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflectioncited by 1,800
  5. 2023FActScore: Fine-grained Atomic Evaluation of Factual Precision in Long Form Text Generationcited by 1,259
  6. 2022When Not to Trust Language Models: Investigating Effectiveness of Parametric and Non-Parametric Memoriescited by 1,103
  7. 2023On-Policy Distillation of Language Models: Learning from Self-Generated Mistakescited by 840
  8. 2023AlpacaFarm: A Simulation Framework for Methods that Learn from Human Feedbackcited by 738
  9. 2019ELI5: Long Form Question Answeringcited by 567
  10. 2023LMSYS-Chat-1M: A Large-Scale Real-World LLM Conversation Datasetcited by 449
  11. 2024Fine-grained Hallucination Detection and Editing for Language Modelscited by 146
  12. 2024Lookback Lens: Detecting and Mitigating Contextual Hallucinations in Large Language Models Using Only Attention Mapscited by 140
20 more
  1. 2024VERISCORE: Evaluating the factuality of verifiable claims in long-form text generationcited by 114
  2. 2023R-Tuning: Instructing Large Language Models to Say `I Don't Know'cited by 106
  3. 2024Refuse Whenever You Feel Unsafe: Improving Safety in LLMs via Decoupled Refusal Trainingcited by 67
  4. 2024Learning to Route LLMs with Confidence Tokenscited by 66
  5. 2025ToolHop: A Query-Driven Benchmark for Evaluating Large Language Models in Multi-Hop Tool Usecited by 48
  6. 2024Can AI Assistants Know What They Don't Know?cited by 47
  7. 2024FLAME: Factuality-Aware Alignment for Large Language Modelscited by 46
  8. 2024HalluDial: A Large-Scale Benchmark for Automatic Dialogue-Level Hallucination Evaluationcited by 26
  9. 2024FactBench: A Dynamic Benchmark for In-the-Wild Language Model Factuality Evaluationcited by 23
  10. 2025Learning to Reason for Factualitycited by 22
  11. 2025KnowRL: Exploring Knowledgeable Reinforcement Learning for Factualitycited by 17
  12. 2024Improving Model Factuality with Fine-grained Critique-based Evaluatorcited by 15
  13. 2025CompileAgent: Automated Real-World Repo-Level Compilation with Tool-Integrated LLM-based Agent Systemcited by 15
  14. 2024UAlign: Leveraging Uncertainty Estimations for Factuality Alignment on Large Language Modelscited by 13
  15. 2025Mitigating LLM Hallucination via Behaviorally Calibrated Reinforcement Learningcited by 13
  16. 2025FACTORY: A Challenging Human-Verified Prompt Set for Long-Form Factualitycited by 5
  17. 2025Train for Truth, Keep the Skills: Binary Retrieval-Augmented Reward Mitigates Hallucinationscited by 5
  18. 2025Odysseus Navigates the Sirens' Song: Dynamic Focus Decoding for Factual and Diverse Open-Ended Text Generationcited by 4
  19. 2025Teaching LLMs to Abstain via Fine-Grained Semantic Confidence Rewardcited by 2
  20. 2026Two Pathways to Truthfulness: On the Intrinsic Encoding of LLM Hallucinationscited by 1

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Think Through Uncertainty: Improving Long-Form Generation Factuality via Reasoning Calibration
  2. 2026Calibration Drift Under Reasoning: How Chain-of-Thought Budgets Induce Overconfidence in Large Language Models
  3. 2025MR-Align: Meta-Reasoning Informed Factuality Alignment for Large Reasoning Models
  4. 2025Thought calibration: Efficient and confident test-time scaling
  5. 2025Trustworthy Reasoning: Evaluating and Enhancing Factual Accuracy in LLM Intermediate Thought Processes
  6. 2026EpiCaR: Knowing What You Don't Know Matters for Better Reasoning in LLMs
  7. 2026On the Pitfalls of Verbalized Confidence Priors for Calibrating Large Reasoning Models
  8. 2025MONICA: Real-Time Monitoring and Calibration of Chain-of-Thought Sycophancy in Large Reasoning Models