Lune

ACL Findings2026

Robust Uncertainty Quantification for Self-Evolving Large Language Models via Continual Domain Pretraining

Xiaofan Zhou, Lu Cheng

Abstract

Continual Learning (CL) is essential for enabling self-evolving large language models (LLMs) to adapt and remain effective amid rapid knowledge growth. Yet, despite its importance, little attention has been given to establishing statistical reliability guarantees for LLMs under CL, particularly in the setting of continual domain pretraining (CDP). Conformal Prediction (CP) has shown promise in offering correctness guarantees for LLMs, but it faces major challenges in CDP: testing data often stems from unknown or shifting domain distributions, under which CP may no longer provide valid guarantees. Moreover, when high coverage is required, CP can yield excessively large prediction sets for unanswerable queries, reducing informativeness. To address these challenges, we introduce an adaptive rejection and non-exchangeable CP framework. Our method first estimates the distribution of questions across domains in the test set using transformer-based clustering, then reweights or resamples the calibration data accordingly. Building on this, adaptive rejection CP allows the LLM to selectively abstain from answering when its confidence or competence shifts significantly. Extensive experiments demonstrate that our framework enhances both the effectiveness and reliability of CP under CDP scenarios. Our code is available at: https://github.com/AlearZhou/CPCL 20 40 60 80 Quantile (%) 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Value Quantile Distribution (Normalized Entropy) sc ht sp en gg whole 20 40 60 80 Quantile (%)

Ask Lune

Know this paper in minutes, not an afternoon.

Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.

1 top-tier paper has cited it since. Know what they found before your reviewers do.

Questions to start from

Lune can draw on

the full paper · the top-tier papers that cite it · related work

Get the answer

Free to start. No credit card required.

Lineage

  • Conformal Alignment: Knowing When to Trust Foundation Models with Guarantees cites Conformal Language Modeling
  • Conformal Alignment: Knowing When to Trust Foundation Models with Guarantees cites Language Models with Conformal Factuality Guarantees
  • ELLE: Efficient Lifelong Pre-training for Emerging Data cites Don't Stop Pretraining: Adapt Language Models to Domains and Tasks
  • Language Models with Conformal Factuality Guarantees cites Conformal Language Modeling
Builds onCited byLeft to right by year; the more cited, the higher.

Cited by top-tier papers1

  1. AIM-CoT: Active Information-driven Multimodal Chain-of-Thought for Vision-Language ReasoningXiping Li, Jianghong MaACL 2026 · 3 citations

Builds on15

  1. Conformal Language ModelingVictor Quach, Adam Fisch, Tal Schuster, Adam Yala et al.ICLR 2024 · 132 citations
  2. Language Models with Conformal Factuality GuaranteesChristopher Mohri, Tatsunori HashimotoICML 2024 · 107 citations
  3. Don't Stop Pretraining: Adapt Language Models to Domains and TasksSuchin Gururangan, Ana Marasovic, Swabha Swayamdipta, Kyle Lo et al.ACL 2020 · 93 citations
  4. Conformal Alignment: Knowing When to Trust Foundation Models with GuaranteesYu Gui, Ying Jin, Zhimei RenNeurIPS 2024 · 63 citations
  5. ELLE: Efficient Lifelong Pre-training for Emerging DataYujia Qin, Jiajie Zhang, Yankai Lin, Zhiyuan Liu et al.ACL Findings 2022 · 32 citations

Related papers

  1. Model Uncertainty Quantification by Conformal Prediction in Continual LearningRui Gao, Weiwei LiuICML 2025
  2. Domain-Shift-Aware Conformal Prediction for Large Language ModelsZhexiao Lin, Yuanyuan Li, Neeraj Sarna, Yuanyuan Gao et al.ICML 2026 · 6 citations
  3. Prune 'n Predict: Optimizing LLM Decision-making with Conformal PredictionHarit Vishwakarma, Alan Mishler, Thomas Cook, Niccolò Dalmasso et al.ICML 2025
  4. ConU: Conformal Uncertainty in Large Language Models with Correctness Coverage GuaranteesZhiyuan Wang, Jinhao Duan, Lu Cheng, Yue Zhang et al.EMNLP Findings 2024 · 16 citations
  5. Conformal Linguistic Calibration: Trading-off between Factuality and SpecificityZhengping Jiang, Anqi Liu, Benjamin Van DurmeNeurIPS 2025 · 22 citations
Dusk over the sea between two cliffs drawn in fine vertical lines