Introduces unified span-level hallucination detection benchmark over code, tool output, and documents; fine-tuned Qwen3.5-2B reaches 0.689 span-F1 and outperforms baselines including on code-agent data.
hub
HaluEval: A Large-Scale Hallucination Evaluation Benchmark for Large Language Models
23 Pith papers cite this work, alongside 199 external citations. Polarity classification is still indexing.
23
Pith papers citing it
199
external citations · Crossref
hub tools
citation-role summary
background 1
citation-polarity summary
roles
background 1polarities
background 1representative citing papers
An empirical study distills a taxonomy of human factual errors from newspaper corrections and shows LLMs achieve only 52% F1 on detection.
AuthorityBench shows citation presence (real or fabricated) increases LLM hallucination rates vs no-citation baseline, strongest for fabricated citations on true claims, with domain variation but negligible venue or author effects.
PhantomBench is a new benchmark of 60K+ non-existent terms showing language models hallucinate at rates up to 86.7 percent even when inputs assume the concepts exist.
SuperMemory-VQA provides 4,853 human-verified QA pairs from 52.9 hours of egocentric AI glasses recordings to benchmark AI systems on realistic long-horizon memory tasks including an unanswerable option.
A three-axis taxonomy of LLM deception, applied to 50 benchmarks, shows fabrication is heavily covered while omission, pragmatic distortion, attribution, and capability self-knowledge remain gaps.
Hallucinations arise from biased latent inference paths rather than missing knowledge, demonstrated via a new diagnostic testbed TrapQA that isolates task-retrieval and key-selection biases.
The paper characterizes deductive stereotyping in LLMs and introduces Fair-GCG to discover injection phrases that improve fairness across benchmarks, reasoning, and real-world tasks.
Global calibration metrics like ECE are confounded by accuracy; the proposed ACE framework with three accuracy-controlled views shows many prior calibration advantages weaken or reverse.
AURORA detects hallucinations via skewness of cosine similarities between weights and gradients plus a rotation ratio from SVD on update-induced changes to singular vectors.
TriLens detects hallucinations via per-layer entropy trajectories of logit-lens readouts from three internal modules across LLMs and QA benchmarks.
A GNN trained on bipartite alignment graphs between references and LLM generations reports state-of-the-art hallucination detection across four datasets, beating prior methods and GPT-4o.
TRACE uses cross-layer candidate trajectories inside frozen LLMs to dynamically select and apply one of three correction operators, delivering mean gains of +12.26 MC1 and +8.65 MC2 points across 15 models and 3 benchmarks with no regressions.
LaaB improves LLM hallucination detection by mapping self-judgment labels back into neural feature space and using mutual learning under logical consistency constraints between responses and meta-judgments.
Adaptive Unlearning suppresses package hallucinations in code-generating LLMs by 81% while preserving benchmark performance, using model-generated data and no human labels.
Know2Guess is a contamination-aware multi-zone benchmark for evaluating LLM knowledge boundaries with explicit abstention expectations and dual parsers.
LAG-XAI treats paraphrasing as affine flows in semantic manifolds using Lie-inspired approximations, achieving AUC 0.7713 on paraphrase detection and 95.3% hallucination detection on HaluEval.
ReFACT benchmark reveals LLMs show a persistent salient distractor failure mode where 61% of incorrect error span predictions are semantically unrelated to actual errors, persisting across model sizes, and comparative judgment yields lower F1 than independent detection.
CORTEX detects token-level hallucinations in RAG via comparative internal representations, information propagation, and smoothing, reporting gains on two benchmarks with three LLMs.
Introduces LLM-FACETS, a privacy-preserving open-source framework for LLM evaluation using deterministic metrics locally, LLM-judge metrics with user-controlled APIs, and mechanisms for uncertainty visualization and hallucination detection.
Any spectral diagnostic that depends only on singular values or the symmetric part of a degree-normalized attention matrix is invariant under transpose, so it cannot see the direction of information flow.
EPGS detects high-confidence factual errors in LLMs by using embedding perturbations to measure gradient sensitivity as a proxy for sharp versus flat minima.
A thermodynamic-inspired information-geometric framework defines a composite LLM stability score that outperforms a utility-entropy baseline by 0.0299 on average across 80 observations, with gains increasing at higher entropy.
citing papers explorer
-
Beyond Document Grounding: Span-Level Hallucination Detection over Code, Tool Output, and Documents
Introduces unified span-level hallucination detection benchmark over code, tool output, and documents; fine-tuned Qwen3.5-2B reaches 0.689 span-F1 and outperforms baselines including on code-agent data.
-
An Empirical Analysis of Factual Errors in Human-Written Text and its Application
An empirical study distills a taxonomy of human factual errors from newspaper corrections and shows LLMs achieve only 52% F1 on detection.
-
Authority, Truth, and Citation Bias: A Large-Scale Multi-Domain Benchmark for Studying Epistemic Susceptibility in Large Language Models
AuthorityBench shows citation presence (real or fabricated) increases LLM hallucination rates vs no-citation baseline, strongest for fabricated citations on true claims, with domain variation but negligible venue or author effects.
-
PhantomBench: Benchmarking the Non-existential Threat of Language Models
PhantomBench is a new benchmark of 60K+ non-existent terms showing language models hallucinate at rates up to 86.7 percent even when inputs assume the concepts exist.
-
SuperMemory-VQA: An Egocentric Visual Question-Answering Benchmark for Long-Horizon Memory
SuperMemory-VQA provides 4,853 human-verified QA pairs from 52.9 hours of egocentric AI glasses recordings to benchmark AI systems on realistic long-horizon memory tasks including an unanswerable option.
-
From Sycophancy to Deception: A Unified Taxonomy for LLM Spontaneous Misalignment
A three-axis taxonomy of LLM deception, applied to 50 benchmarks, shows fabrication is heavily covered while omission, pragmatic distortion, attribution, and capability self-knowledge remain gaps.
-
Understanding Why Language Models Hallucinate: Testing Reasoning Against Priors
Hallucinations arise from biased latent inference paths rather than missing knowledge, demonstrated via a new diagnostic testbed TrapQA that isolates task-retrieval and key-selection biases.
-
Wait, am I Being Fair? Characterizing Deductive Stereotyping and Mitigating It with Fair-GCG
The paper characterizes deductive stereotyping in LLMs and introduces Fair-GCG to discover injection phrases that improve fairness across benchmarks, reasoning, and real-world tasks.
-
When Calibration Rankings Reverse: Accuracy-Controlled Evaluation for Fair Comparison of LLMs
Global calibration metrics like ECE are confounded by accuracy; the proposed ACE framework with three accuracy-controlled views shows many prior calibration advantages weaken or reverse.
-
AURORA: Asymmetry and Update-Induced Rotation for Robust Hallucination Detection in Large Language Models
AURORA detects hallucinations via skewness of cosine similarities between weights and gradients plus a rotation ratio from SVD on update-induced changes to singular vectors.
-
TriLens: Per-Layer Logit-Lens Entropy for White-Box Hallucination Detection
TriLens detects hallucinations via per-layer entropy trajectories of logit-lens readouts from three internal modules across LLMs and QA benchmarks.
-
Graph Alignment Topology as an Inductive Bias for Grounding Detection
A GNN trained on bipartite alignment graphs between references and LLM generations reports state-of-the-art hallucination detection across four datasets, beating prior methods and GPT-4o.
-
TRACE: Trajectory Correction from Cross-layer Evidence for Hallucination Reduction
TRACE uses cross-layer candidate trajectories inside frozen LLMs to dynamically select and apply one of three correction operators, delivering mean gains of +12.26 MC1 and +8.65 MC2 points across 15 models and 3 benchmarks with no regressions.
-
Logical Consistency as a Bridge: Improving LLM Hallucination Detection via Label Constraint Modeling between Responses and Self-Judgments
LaaB improves LLM hallucination detection by mapping self-judgment labels back into neural feature space and using mutual learning under logical consistency constraints between responses and meta-judgments.
-
LLM Ghostbusters: Surgical Hallucination Suppression via Adaptive Unlearning
Adaptive Unlearning suppresses package hallucinations in code-generating LLMs by 81% while preserving benchmark performance, using model-generated data and no human labels.
-
Know2Guess: A Contamination-Aware Multi-Zone Benchmark for Knowledge-Boundary Evaluation in Large Language Models
Know2Guess is a contamination-aware multi-zone benchmark for evaluating LLM knowledge boundaries with explicit abstention expectations and dual parsers.
-
LAG-XAI: A Lie-Inspired Affine Geometric Framework for Interpretable Paraphrasing in Transformer Latent Spaces
LAG-XAI treats paraphrasing as affine flows in semantic manifolds using Lie-inspired approximations, achieving AUC 0.7713 on paraphrase detection and 95.3% hallucination detection on HaluEval.
-
ReFACT: A Benchmark for Scientific Confabulation Detection with Positional Error Annotations
ReFACT benchmark reveals LLMs show a persistent salient distractor failure mode where 61% of incorrect error span predictions are semantically unrelated to actual errors, persisting across model sizes, and comparative judgment yields lower F1 than independent detection.
-
CORTEX: Token-Level Hallucination Detection in RAG via Comparative Internal Representations
CORTEX detects token-level hallucinations in RAG via comparative internal representations, information propagation, and smoothing, reporting gains on two benchmarks with three LLMs.
-
LLM-FACETS: A Privacy-Preserving Framework for Evaluating LLM Transparency and Accountability
Introduces LLM-FACETS, a privacy-preserving open-source framework for LLM evaluation using deterministic metrics locally, LLM-judge metrics with user-controlled APIs, and mechanisms for uncertainty visualization and hallucination detection.
-
Self-Attention as Transport: Limits of Symmetric Spectral Diagnostics
Any spectral diagnostic that depends only on singular values or the symmetric part of a degree-normalized attention matrix is invariant under transpose, so it cannot see the direction of information flow.
-
From Flat Facts to Sharp Hallucinations: Detecting Stubborn Errors via Gradient Sensitivity
EPGS detects high-confidence factual errors in LLMs by using embedding perturbations to measure gradient sensitivity as a proxy for sharp versus flat minima.
-
An Information-Geometric Framework for Stability Analysis of Large Language Models under Entropic Stress
A thermodynamic-inspired information-geometric framework defines a composite LLM stability score that outperforms a utility-entropy baseline by 0.0299 on average across 80 observations, with gains increasing at higher entropy.