From Tokens to Semantics: Leveraging Complementary Signals for Hallucination Detection in Black-Box LLMs

• URL: http://arxiv.org/abs/2609.02679v3
• Date: Mon, 07 Sep 2026 09:22:26 GMT
• Title: From Tokens to Semantics: Leveraging Complementary Signals for Hallucination Detection in Black-Box LLMs
• Abstract summary: Missing fabrications can harm users and institutions, while false alarms consume limited human-review capacity.<n>We study two signals accessible through black-box model APIs: semantic entropy, and uncertainty derived from token log-probabilities.<n>We extend token-based uncertainty detection by aggregating token-level signals across sampled responses through our TopK method.
• Score: 0.9751538234290761
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: When LLMs support public-facing or high-stakes workflows, missed fabrications can harm users and institutions, while false alarms consume limited human-review capacity. When no trusted context or reference document is available, we study two signals accessible through black-box model APIs: semantic entropy, which measures disagreement among sampled response meanings, and uncertainty derived from token log-probabilities. Their failure modes can be complementary: semantic entropy becomes uninformative when responses form one semantic cluster, while token uncertainty can miss consistently confident errors. We extend token-based uncertainty detection by aggregating token-level signals across sampled responses through our TopK method, evaluate the hybrid CoCoA method, which combines target-response uncertainty with semantic dissimilarity, and propose and study two supervised methods: Gated, which routes single-cluster cases to an aggregated-token-feature classifier, and Stacked, which learns jointly from semantic uncertainty and broader token features. We evaluate seven benchmarks, including five public benchmarks (four text datasets and multimodal handwritten-cheque extraction) and two constructed benchmarks (Financial Summaries and Long-Text QA), using four language models. In our evaluation across models and datasets, Stacked gave the best performance in nearly half of the cases, while TopK and CoCoA remain competitive without supervised training labels, although their thresholds require careful calibration. No method is universally strongest. We therefore evaluate performance at false-positive-rate budgets from 1% to 15%, assess their sensitivity to generation and calibration choices, and examine variation across dataset characteristics.

Related papers

• Certifying Adversarial Robustness of Quantum Classifiers under Known-Readout Query Access [6.902322288643561]
  A quantum classifier assigns labels by evolving an input quantum state and measuring the output, so repeated executions reveal only a distribution over labels.<n>We give a measurement-only framework that returns two complementary guarantees for each input.<n>From 40 executions of two 8-qubit quantum neural networks, our method estimates both bounds, with the expected lower-upper ordering on every tested input.
  arXiv  Detail & Related papers  (2026-09-10T14:45:21Z)
• Clustered Self-Assessment: A Simple yet Effective Method for Uncertainty Quantification in Large Language Models [44.59799077650502]
  We propose a simple yet effective self-assessment method for uncertainty quantification in large language models.<n>Our approach groups sampled generations into semantically distinct clusters, converts them into answer options in a structured multiple-choice question, and uses the probability assigned by the model to each option as a confidence estimate.
  arXiv  Detail & Related papers  (2026-06-02T16:25:54Z)
• 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)
• Know2Guess: A Contamination-Aware Multi-Zone Benchmark for Knowledge-Boundary Evaluation in Large Language Models [9.503036150359597]
  We present a contamination-aware, multi-zone benchmark for measuring the transition from answerable knowledge to abstention-expected unknowns.<n>We evaluate FLAN-T5, Qwen2.5-Instruct, and Llama-3-Instruct models under locked answer-or-abstain prompts, answer-only controls, and prompt-template variants.
  arXiv  Detail & Related papers  (2026-04-30T05:46:01Z)
• Beyond Raw Detection Scores: Markov-Informed Calibration for Boosting Machine-Generated Text Detection [105.14032334647932]
  Machine-generated texts (MGTs) pose risks such as disinformation and phishing, highlighting the need for reliable detection.<n> Metric-based methods, which extract statistically distinguishable features of MGTs, are often more practical than complex model-based methods that are prone to overfitting.<n>We propose a Markov-informed score calibration strategy that models two relationships of context detection scores that may aid calibration.
  arXiv  Detail & Related papers  (2026-02-08T16:06:12Z)
• CLUE: Non-parametric Verification from Experience via Hidden-State Clustering [64.50919789875233]
  We show that correctness of a solution is encoded as a geometrically separable signature within the trajectory of hidden activations.<n>ClUE consistently outperforms LLM-as-a-judge baselines and matches or exceeds modern confidence-based methods in reranking candidates.
  arXiv  Detail & Related papers  (2025-10-02T02:14:33Z)
• Reference-Free Rating of LLM Responses via Latent Information [53.463883683503106]
  We study the common practice of asking a judge model to assign Likert-scale scores to free-text responses.<n>We then propose and evaluate Latent Judges, which derive scalar ratings from internal model signals.<n>Across a broad suite of pairwise and single-rating benchmarks, latent methods match or surpass standard prompting.
  arXiv  Detail & Related papers  (2025-09-29T12:15:52Z)
• Random-Set Large Language Models [4.308457163593758]
  Large Language Models (LLMs) are known to produce very high-quality tests and responses to our queries.<n>But how much can we trust this generated text?<n>We propose a novel Random-Set Large Language Model (RSLLM) approach which predicts finite random sets (belief functions) over the token space.
  arXiv  Detail & Related papers  (2025-04-25T05:25:27Z)
• Pretraining Data Detection for Large Language Models: A Divergence-based Calibration Method [108.56493934296687]
  We introduce a divergence-based calibration method, inspired by the divergence-from-randomness concept, to calibrate token probabilities for pretraining data detection.<n>We have developed a Chinese-language benchmark, PatentMIA, to assess the performance of detection approaches for LLMs on Chinese text.
  arXiv  Detail & Related papers  (2024-09-23T07:55:35Z)
• Token-Level Adversarial Prompt Detection Based on Perplexity Measures and Contextual Information [67.78183175605761]
  Large Language Models are susceptible to adversarial prompt attacks. This vulnerability underscores a significant concern regarding the robustness and reliability of LLMs. We introduce a novel approach to detecting adversarial prompts at a token level.
  arXiv  Detail & Related papers  (2023-11-20T03:17:21Z)
• Exploiting Sample Uncertainty for Domain Adaptive Person Re-Identification [137.9939571408506]
  We estimate and exploit the credibility of the assigned pseudo-label of each sample to alleviate the influence of noisy labels. Our uncertainty-guided optimization brings significant improvement and achieves the state-of-the-art performance on benchmark datasets.
  arXiv  Detail & Related papers  (2020-12-16T04:09:04Z)

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.