paper-with-me

홈 › Papers

CATTO: Balancing Preferences and Confidence in Language Models

2026-01-30 · Nisarg Parikh, Ananya Sai, Pannaga Shivaswamy, Kunjal Panchal, Andrew Lan arxiv

Large language models (LLMs) often make accurate next token predictions but their confidence in these predictions can be poorly calibrated: high-confidence predictions are frequently wrong, and low-confidence predictions may be correct. This miscalibration is exacerbated by preference-based alignment methods breaking the link between predictive probability and correctness. We introduce a Calibration Aware Token-level Training Objective (CATTO), a calibration-aware objective that aligns predicted confidence with empirical prediction correctness, which can be combined with the original preference optimization objectives. Empirically, CATTO reduces Expected Calibration Error (ECE) by 2.22%-7.61% in-distribution and 1.46%-10.44% out-of-distribution compared to direct preference optimization (DPO), and by 0.22%-1.24% in-distribution and 1.23%-5.07% out-of-distribution compared to the strongest DPO baseline. This improvement in confidence does not come at a cost of losing task accuracy, where CATTO maintains or slightly improves multiple-choice question-answering accuracy on five datasets. We also introduce Confidence@k, a test-time scaling mechanism leveraging calibrated token probabilities for Bayes-optimal selection of output tokens.

📄 PDF Abstract BibTeX arXiv:2601.23096

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Vehicle Rebalancing Under Adherence Uncertainty

2024-12-21 · Avalpreet Singh Brar, Rong Su, Gioele Zardini

Ride hailing systems suffer from spatial-temporal supply demand imbalance due to drivers operating in independent, and uncoordinated manner. Several fleet rebalancing models have been proposed that can provide reposition…

CheXPO: Preference Optimization for Chest X-ray VLMs with Counterfactual Rationale

2025-07-09 · Xiao Liang, Jiawei Hu, Di Wang, Zhi Ma 외 arxiv

Vision-language models (VLMs) are prone to hallucinations that critically compromise reliability in medical applications. While preference optimization can mitigate these hallucinations through clinical feedback, its imp…

CAPO: Confidence Aware Preference Optimization Learning for Multilingual Preferences

2025-11-10 · Rhitabrat Pokharel, Yufei Tao, Ameeta Agrawal arxiv

Preference optimization is a critical post-training technique used to align large language models (LLMs) with human preferences, typically by fine-tuning on ranked response pairs. While methods like Direct Preference Opt…

Leveraging Large Language Models for Collective Decision-Making

2023-11-03 · Marios Papachristou, Longqi Yang, Chin-Chia Hsu

In various work contexts, such as meeting scheduling, collaborating, and project planning, collective decision-making is essential but often challenging due to diverse individual preferences, varying work focuses, and po…

Decision MakingScheduling

CoPL: Collaborative Preference Learning for Personalizing LLMs

2025-03-03 · Youngbin Choi, Seunghyuk Cho, Minjong Lee, Moonjeong Park 외

Personalizing large language models (LLMs) is important for aligning outputs with diverse user preferences, yet existing methods struggle with flexibility and generalization. We propose CoPL (Collaborative Preference Lea…

Collaborative Filtering