Entropy Alone is Insufficient for Safe Selective Prediction in LLMs

• URL: http://arxiv.org/abs/2603.21172v1
• Date: Sun, 22 Mar 2026 11:27:13 GMT
• Title: Entropy Alone is Insufficient for Safe Selective Prediction in LLMs
• Abstract summary: Selective prediction systems can mitigate harms resulting from language model hallucinations by abstaining from answering in high-risk cases.<n>Uncertainty quantification techniques are often employed to identify such cases, but are rarely evaluated in the context of the wider selective prediction policy.<n>We identify a model-dependent failure mode of entropy-based uncertainty methods that leads to unreliable abstention behaviour, and address it by combining entropy scores with a correctness probe signal.
• Score: 20.664633053172327
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Selective prediction systems can mitigate harms resulting from language model hallucinations by abstaining from answering in high-risk cases. Uncertainty quantification techniques are often employed to identify such cases, but are rarely evaluated in the context of the wider selective prediction policy and its ability to operate at low target error rates. We identify a model-dependent failure mode of entropy-based uncertainty methods that leads to unreliable abstention behaviour, and address it by combining entropy scores with a correctness probe signal. We find that across three QA benchmarks (TriviaQA, BioASQ, MedicalQA) and four model families, the combined score generally improves both the risk--coverage trade-off and calibration performance relative to entropy-only baselines. Our results highlight the importance of deployment-facing evaluation of uncertainty methods, using metrics that directly reflect whether a system can be trusted to operate at a stated risk level.

Related papers

• Asymptotic Risk Calibration for Selective Question Answering [0.0]
  Large language models (LLMs) may generate fluent but incorrect answers.<n>We propose A-CRC-QA, a post-hoc calibration framework for uncertainty-aware selective question answering.<n>A-CRC-QA is model-agnostic, requires no additional training, and can be combined with different uncertainty estimators.
  arXiv  Detail & Related papers  (2026-08-12T12:45:45Z)
• Uncertainty-Aware Abstention in Large Language Models with Provable Alignment Guarantees [0.5371337604556311]
  Large language models (LLMs) are increasingly deployed in question answering (QA) systems.<n>They may generate hallucinated or misaligned responses without reliable confidence estimates.<n>We propose CIC, a confidence-interval-based calibration framework that converts arbitrary uncertainty scores into risk-controlled selective answering rules.
  arXiv  Detail & Related papers  (2026-07-05T17:50:32Z)
• On the QUEST for Uncertainty Quantification via Highest Density Regions [5.1459761018260854]
  Uncertainty quantification (UQ) is essential for reliable decision-making in safety-critical applications in machine learning.<n>We propose an alternative framework, in which uncertainty is characterised by the volume of the most probable subset of a distribution's support.<n>QUEST (Quantifying Uncertainty via highest dEnSiTy regions) is a novel approach to UQ based on the concentration of Lebesgue measure at a distribution's peak(s)
  arXiv  Detail & Related papers  (2026-06-17T20:13:45Z)
• LLMs Uncertainty Quantification via Adaptive Conformal Semantic Entropy [1.3254304182988286]
  We propose Adaptive Conformal Semantic Entropy (ACSE), a method for estimating prompt-level uncertainty.<n>Our uncertainty scoring function is based on clustering semantic entropy of multiple diverse responses to the same prompt.<n>Our approach consistently outperforms state-of-the-art uncertainty quantification baselines.
  arXiv  Detail & Related papers  (2026-05-05T20:56:11Z)
• LEC: Linear Expectation Constraints for False-Discovery Control in Selective Prediction and Routing Systems [95.35293543918762]
  Large language models (LLMs) often generate unreliable answers, while uncertainty methods fail to fully distinguish correct from incorrect predictions.<n>We address this issue through the lens of false discovery rate (FDR) control, ensuring that among all accepted predictions, the proportion of errors does not exceed a target risk level.<n>We propose LEC, which reinterprets selective prediction as a constrained decision problem by enforcing a Linear Expectation Constraint.
  arXiv  Detail & Related papers  (2025-12-01T11:27:09Z)
• Addressing Pitfalls in the Evaluation of Uncertainty Estimation Methods for Natural Language Generation [20.726685669562496]
  Hallucinations are a common issue that undermine the reliability of large language models (LLMs)<n>Recent studies have identified a subset of hallucinations, known as confabulations, which arise due to predictive uncertainty of LLMs.<n>To detect confabulations, various methods for estimating predictive uncertainty in natural language generation (NLG) have been developed.
  arXiv  Detail & Related papers  (2025-10-02T17:54:09Z)
• Uncertainty Quantification for Regression using Proper Scoring Rules [76.24649098854219]
  We introduce a unified UQ framework for regression based on proper scoring rules, such as CRPS, logarithmic, squared error, and quadratic scores.<n>We derive closed-form expressions for the uncertainty measures under practical parametric assumptions and show how to estimate them using ensembles of models.<n>Our broad evaluation on synthetic and real-world regression datasets provides guidance for selecting reliable UQ measures.
  arXiv  Detail & Related papers  (2025-09-30T17:52:12Z)
• COIN: Uncertainty-Guarding Selective Question Answering for Foundation Models with Provable Risk Guarantees [51.5976496056012]
  COIN is an uncertainty-guarding selection framework that calibrates statistically valid thresholds to filter a single generated answer per question.<n>COIN estimates the empirical error rate on a calibration set and applies confidence interval methods to establish a high-probability upper bound on the true error rate.<n>We demonstrate COIN's robustness in risk control, strong test-time power in retaining admissible answers, and predictive efficiency under limited calibration data.
  arXiv  Detail & Related papers  (2025-06-25T07:04:49Z)
• SConU: Selective Conformal Uncertainty in Large Language Models [59.25881667640868]
  We propose a novel approach termed Selective Conformal Uncertainty (SConU)<n>We develop two conformal p-values that are instrumental in determining whether a given sample deviates from the uncertainty distribution of the calibration set at a specific manageable risk level.<n>Our approach not only facilitates rigorous management of miscoverage rates across both single-domain and interdisciplinary contexts, but also enhances the efficiency of predictions.
  arXiv  Detail & Related papers  (2025-04-19T03:01:45Z)
• Label-Confidence-Aware Uncertainty Estimation in Natural Language Generation [8.635811152610604]
  Uncertainty Quantification (UQ) is crucial for ensuring the safety and robustness of AI systems.<n>We propose a label-confidence-aware (LCA) uncertainty estimation based on Kullback-Leibler divergence bridging between samples and label source.
  arXiv  Detail & Related papers  (2024-12-10T07:35:23Z)
• Mitigating optimistic bias in entropic risk estimation and optimization with an application to insurance [5.407319151576265]
  The entropic risk measure is widely used to account for tail risks associated with an uncertain loss.<n>To mitigate the bias in the empirical entropic risk estimator, we propose a strongly consistent bootstrapping procedure.<n>We show that our methods suggest a higher (and more accurate) premium to homeowners.
  arXiv  Detail & Related papers  (2024-09-30T04:02:52Z)
• Model-Based Epistemic Variance of Values for Risk-Aware Policy Optimization [59.758009422067]
  We consider the problem of quantifying uncertainty over expected cumulative rewards in model-based reinforcement learning. We propose a new uncertainty Bellman equation (UBE) whose solution converges to the true posterior variance over values. We introduce a general-purpose policy optimization algorithm, Q-Uncertainty Soft Actor-Critic (QU-SAC) that can be applied for either risk-seeking or risk-averse policy optimization.
  arXiv  Detail & Related papers  (2023-12-07T15:55:58Z)
• DEUP: Direct Epistemic Uncertainty Prediction [56.087230230128185]
  Epistemic uncertainty is part of out-of-sample prediction error due to the lack of knowledge of the learner. We propose a principled approach for directly estimating epistemic uncertainty by learning to predict generalization error and subtracting an estimate of aleatoric uncertainty.
  arXiv  Detail & Related papers  (2021-02-16T23:50:35Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.