A DETR-style probe distills multi-sample claim uncertainty into single-pass span detection and continuous Mixture-of-Beta scores, outperforming baselines on a new 293K-span benchmark.
TokUR: Token-Level Uncertainty Estimation for Large Language Model Reasoning
6 Pith papers cite this work. Polarity classification is still indexing.
6
Pith papers citing it
abstract
While Large Language Models (LLMs) have demonstrated impressive capabilities, their output quality remains inconsistent across various application scenarios, making it difficult to identify trustworthy responses, especially in complex tasks requiring multi-step reasoning. In this paper, we propose a Token-level Uncertainty estimation framework for Reasoning (TokUR) that enables LLMs to self-assess and self-improve their responses in mathematical reasoning. Specifically, we introduce low-rank random weight perturbation during LLM decoding to generate predictive distributions for token-level uncertainty estimation, and we aggregate these uncertainty quantities to capture the semantic uncertainty of generated responses. Experiments on mathematical reasoning datasets of varying difficulty demonstrate that TokUR exhibits a strong correlation with answer correctness and model robustness, and the uncertainty signals produced by TokUR can be leveraged to enhance the model's reasoning performance at test time. These results highlight the effectiveness of TokUR as a principled and scalable approach for improving the reliability and interpretability of LLMs in challenging reasoning tasks.
citation-role summary
background 1
citation-polarity summary
years
2026 6roles
background 1polarities
background 1representative citing papers
LLMs exhibit epistemically vacuous confidence on clinical tabular data, but cross-model attribution divergence with XGBoost enables a calibrator that reduces expected calibration error from 0.254 to 0.080.
BAS aggregates utility from an answer-or-abstain model across risk thresholds and is uniquely maximized by truthful confidence estimates.
SFT supplies entangled compositional traces of atomic skills and routing modules; RL identifies those modules and enables recombination on novel compositions outside the SFT support.
DPUA is a two-phase framework that aligns LLM uncertainty expressions with human disagreement distributions in subjectivity analysis while preserving task performance.
Erroneous processing heads in attention layers cause hop-generalization failures in LLMs; dynamically deactivating them at test time improves multi-step reasoning.
citing papers explorer
-
SpanUQ: Span-Level Uncertainty Quantification for Large Language Model Generation
A DETR-style probe distills multi-sample claim uncertainty into single-pass span detection and continuous Mixture-of-Beta scores, outperforming baselines on a new 293K-span benchmark.
-
LLM Doesn't Know What It Doesn't Know: Detecting Epistemic Blind Spots via Cross-Model Attribution Divergence on Clinical Tabular Data
LLMs exhibit epistemically vacuous confidence on clinical tabular data, but cross-model attribution divergence with XGBoost enables a calibrator that reduces expected calibration error from 0.254 to 0.080.
-
BAS: A Decision-Theoretic Approach to Evaluating Large Language Model Confidence
BAS aggregates utility from an answer-or-abstain model across risk thresholds and is uniquely maximized by truthful confidence estimates.
-
From Reasoning Traces to Reusable Modules: Understanding Compositional Generalization in Language Model Reasoning
SFT supplies entangled compositional traces of atomic skills and routing modules; RL identifies those modules and enables recombination on novel compositions outside the SFT support.
-
Aligning LLM Uncertainty with Human Disagreement in Subjectivity Analysis
DPUA is a two-phase framework that aligns LLM uncertainty expressions with human disagreement distributions in subjectivity analysis while preserving task performance.
-
Scaling Reasoning Hop Exposes Weaknesses: Demystifying and Improving Hop Generalization in Large Language Models
Erroneous processing heads in attention layers cause hop-generalization failures in LLMs; dynamically deactivating them at test time improves multi-step reasoning.