Conformal Prediction Sets for Next-Token Prediction in Large Language Models: Balancing Coverage Guarantees with Set Efficiency

2025-12-27 · Yoshith Roy Kotla, Varshith Roy Kotla arxiv

Deploying large language models (LLMs) in high-stakes domains requires rigorous uncertainty quantification, yet standard softmax probabilities are often poorly calibrated. We present a systematic study of Adaptive Prediction Sets (APS) applied to next-token prediction in transformer-based models with large vocabularies (greater than 250,000 tokens). Our central contribution is the identification of a coverage-efficiency tradeoff: while naive conformal prediction achieves valid coverage, it produces prediction sets of hundreds of tokens, rendering them uninformative. We propose Vocabulary-Aware Conformal Prediction (VACP), a framework that leverages semantic masking and temperature-adjusted scoring to reduce the effective prediction space while provably maintaining marginal coverage. Experiments on Gemma-2B using SQUAD and WikiText benchmarks demonstrate that VACP achieves 89.7 percent empirical coverage (90 percent target) while reducing the mean prediction set size from 847 tokens to 4.3 tokens -- a 197x improvement in efficiency. We provide a theoretical analysis of vocabulary reduction and release our implementation for reproducibility.

📄 PDF Abstract BibTeX arXiv:2512.22682

Code (0)

등록된 구현이 없습니다.

TECP: Token-Entropy Conformal Prediction for LLMs

2025-08-30 · Beining Xu, Yongming Lu arxiv

Uncertainty quantification (UQ) for open-ended language generation remains a critical yet underexplored challenge, especially under black-box constraints where internal model signals are inaccessible. In this paper, we i…

Provably Robust Conformal Prediction with Improved Efficiency

2024-04-30 · Ge Yan, Yaniv Romano, Tsui-Wei Weng

Conformal prediction is a powerful tool to generate uncertainty sets with guaranteed coverage using any predictive model, under the assumption that the training and test data are i.i.d.. Recently, it has been shown that …

Conformal Prediction Prediction

Efficient and Differentiable Conformal Prediction with General Function Classes

2022-02-22 · ICLR 2022 4 · Yu Bai, Song Mei, Huan Wang, Yingbo Zhou 외

Quantifying the data uncertainty in learning tasks is often done by learning a prediction interval or prediction set of the label given the input. Two commonly desired properties for learned prediction sets are \emph{val…

Conformal Prediction image-classification Image Classification Prediction+2

Non-Exchangeable Conformal Language Generation with Nearest Neighbors

2024-02-01 · Dennis Ulmer, Chrysoula Zerva, André F. T. Martins

Quantifying uncertainty in automatically generated text is important for letting humans check potential hallucinations and making systems more reliable. Conformal prediction is an attractive framework to provide predicti…

Conformal Prediction Language Modeling Language Modelling Machine Translation+2

Conformalized Deep Splines for Optimal and Efficient Prediction Sets

2023-11-01 · Nathaniel Diamant, Ehsan Hajiramezanali, Tommaso Biancalani, Gabriele Scalia

Uncertainty estimation is critical in high-stakes machine learning applications. One effective way to estimate uncertainty is conformal prediction, which can provide predictive inference with statistical coverage guarant…

Conformal Prediction Prediction Prediction Intervals