mlsanityData to 7 Oct 2026

Direct Confidence Alignment: Aligning Verbalized Confidence with Internal Confidence In Large Language Models

arXiv 2512.11998 ↗First posted 12 Dec 2025cs.CLcited by 1

On the map: LLM alignment / RLHF / preference optimisation → LLM preference optimization methods

Abstract (arXiv)

Producing trustworthy and reliable Large Language Models (LLMs) has become increasingly important as their usage becomes more widespread. Calibration seeks to achieve this by improving the alignment between the model's confidence and the actual likelihood of its responses being correct or desirable. However, it has been observed that the internal confidence of a model, derived from token probabilities, is not well aligned with its verbalized confidence, leading to misleading results with different calibration methods. In this paper, we propose Direct Confidence Alignment (DCA), a method using Direct Preference Optimization to align an LLM's verbalized confidence with its internal confidence rather than ground-truth accuracy, enhancing model transparency and reliability by ensuring closer alignment between the two confidence measures. We evaluate DCA across multiple open-weight LLMs on a wide range of datasets. To further assess this alignment, we also introduce three new calibration error-based metrics. Our results show that DCA improves alignment metrics on certain model architectures, reducing inconsistencies in a model's confidence expression. However, we also show that it can be ineffective on others, highlighting the need for more model-aware approaches in the pursuit of more interpretable and trustworthy LLMs.

Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Placed in LLM preference optimization methods, one of 10 directions in LLM alignment / RLHF / preference optimisation. Builds on 23 of its 29 references that are in the map. First posted on arXiv 12 Dec 2025.

Builds on

23 of 29 references.

Closest ancestors · 1

Same direction, newest first

  1. 2023A General Theoretical Paradigm to Understand Learning from Human Preferences18 Oct

Other references · 22

  1. 2024The Llama 3 Herd of Modelscited by 14,192
  2. 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 7,916
  3. 2020Measuring Massive Multitask Language Understandingcited by 7,768
  4. 2017On Calibration of Modern Neural Networkscited by 4,860
  5. 2021TruthfulQA: Measuring How Models Mimic Human Falsehoodscited by 2,919
  6. 2018Can a Suit of Armor Conduct Electricity? A New Dataset for Open Book Question Answeringcited by 2,114
  7. 2024Gemma 2: Improving Open Language Models at a Practical Sizecited by 1,928
  8. 2022Language Models (Mostly) Know What They Knowcited by 1,783
  9. 2021Calibrate Before Use: Improving Few-Shot Performance of Language Modelscited by 1,371
  10. 2023Can LLMs Express Their Uncertainty? An Empirical Evaluation of Confidence Elicitation in LLMscited by 946
  11. 2022Teaching Models to Express Their Uncertainty in Wordscited by 731
  12. 2020How Can We Know When Language Models Know? On the Calibration of Language Models for Question Answeringcited by 439
10 more
  1. 2020Calibration of Pre-trained Transformerscited by 307
  2. 2019Cosmos QA: Machine Reading Comprehension with Contextual Commonsense Reasoningcited by 281
  3. 2023A Survey of Confidence Estimation and Calibration in Large Language Modelscited by 253
  4. 2023A Study of Generative Large Language Model for Medical Research and Healthcarecited by 160
  5. 2023Large Language Models in Law: A Surveycited by 128
  6. 2024When to Trust LLMs: Aligning Confidence with Response Qualitycited by 42
  7. 2024Confidence Under the Hood: An Investigation into the Confidence-Probability Alignment in Large Language Modelscited by 34
  8. 2022On the Calibration of Pre-trained Language Models using Mixup Guided by Area Under the Margin and Saliencycited by 31
  9. 2024Are Large Language Models More Honest in Their Probabilistic or Verbalized Confidence?cited by 23
  10. 2024Self-Consistency Boosts Calibration for Math Reasoningcited by 12

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026Verbalized and Internal Probabilities Are Coupled in Large Language Modelscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2023On the Calibration of Large Language Models and Alignment
  2. 2024Does Alignment Tuning Really Break LLMs' Internal Confidence?
  3. 2024Calibrating the Confidence of Large Language Models by Eliciting Fidelity
  4. 2026ConfidenceBench: Evaluating Confidence Calibration in Large Language Models
  5. 2024Calibrating Long-form Generations from Large Language Models
  6. 2026Instinct vs. Reflection: Unifying Token and Verbalized Confidence in Multimodal Large Models
  7. 2026Unlocking the Pre-Trained Model as a Dual-Alignment Calibrator for Post-Trained LLMs
  8. 2025SDA: Steering-Driven Distribution Alignment for Open LLMs without Fine-Tuning