A between-subjects experiment (N=192) finds that token-level uncertainty increases agreement with LLM answers while relation-level uncertainty reduces external verification in medical decision tasks.
hub
Semantic Entropy Probes: Robust and Cheap Hallucination Detection in LLMs
31 Pith papers cite this work. Polarity classification is still indexing.
31
Pith papers citing it
abstract
We propose semantic entropy probes (SEPs), a cheap and reliable method for uncertainty quantification in Large Language Models (LLMs). Hallucinations, which are plausible-sounding but factually incorrect and arbitrary model generations, present a major challenge to the practical adoption of LLMs. Recent work by Farquhar et al. (2024) proposes semantic entropy (SE), which can detect hallucinations by estimating uncertainty in the space semantic meaning for a set of model generations. However, the 5-to-10-fold increase in computation cost associated with SE computation hinders practical adoption. To address this, we propose SEPs, which directly approximate SE from the hidden states of a single generation. SEPs are simple to train and do not require sampling multiple model generations at test time, reducing the overhead of semantic uncertainty quantification to almost zero. We show that SEPs retain high performance for hallucination detection and generalize better to out-of-distribution data than previous probing methods that directly predict model accuracy. Our results across models and tasks suggest that model hidden states capture SE, and our ablation studies give further insights into the token positions and model layers for which this is the case.
hub tools
citation-role summary
baseline 2
citation-polarity summary
roles
baseline 2polarities
baseline 2representative citing papers
Verification prompting in LVLMs is input-dependent and risk-bearing; RSP selectively triggers it via pre-generation uncertainty to avoid performance degradation on easy cases.
Inducing artificial uncertainty on trivial tasks allows training probes that achieve higher calibration on hard data than standard approaches while retaining performance on easy data.
Attractor basins in transformer hidden states unify conflict and hallucination as basin competition or absence, with geometric margin outperforming entropy for detection and a scaling law governing confident hallucination rates.
Clotho ranks LLM test inputs by failure likelihood using pre-generation hidden states and GMMs, achieving 0.716 ROC-AUC after labeling 5.4% of inputs on average across eight tasks and three models, with transfer to proprietary models.
Grad Detect uses internal gradient patterns from one inference pass to predict LLM hallucinations and abstention, outperforming confidence and sampling baselines on Q&A benchmarks with most signal in the final five layers.
Knot estimates QA model sensitivity to candidate knowledge via subset counterfactual training and latent factor coverage, yielding unit rankings that outperform baselines without extra model calls.
Larger LLMs hallucinate more often despite having the correct concept available because instruction tuning causes probability mass to disperse across alternative surface forms instead of concentrating on one.
Geometric features from per-layer MLP update trajectories fed to a sparse linear probe outperform maximum softmax probability for uncertainty quantification under selective abstention, with gains up to 21 AURC points.
Temporal knowledge drift is encoded as a geometrically orthogonal direction in LLM residual streams, independent of correctness and uncertainty.
Probabilistic circuits detect LLM hallucinations as residual-stream anomalies with up to 99% AUROC and enable dynamic correction that raises truthfulness scores while cutting unnecessary output corruption.
Average token log-probability provides a zero-shot confidence signal for small LLMs that matches supervised baselines in-distribution and outperforms them out-of-distribution, with a new retrieval-conditional variant improving further at lower latency.
LLM token rank-frequency distributions converge to a shared Mandelbrot distribution across models and domains, enabling a microsecond-scale statistical primitive for provenance verification and black-box anomaly triage.
RLCM trains LLMs with a margin-enhanced process reward that widens the gap between correct and incorrect reasoning steps, improving calibration on math, code, logic, and science tasks without hurting accuracy.
Diverse language models converge on similar periodic number features with a two-tier hierarchy of Fourier sparsity and geometric separability, acquired via language co-occurrences or multi-token arithmetic.
Temporal difference calibration aligns uncertainty estimates in vision-language-action models with their value functions for better sequential performance.
Unsupervised single-generation confidence calibration for reasoning LLMs via offline self-consistency proxy distillation outperforms baselines on math and QA tasks and improves selective prediction.
SHADE adaptively combines coverage and spectral signals to estimate semantic alphabet size from few LLM samples, yielding better performance than baselines in low-sample regimes for alphabet estimation and QA error detection.
Cross-model semantic disagreement adds an epistemic uncertainty term that improves total uncertainty estimation over self-consistency alone, helping flag confident errors in LLMs.
Hallucination neurons in LLMs are domain-specific, with cross-domain classifiers dropping from AUROC 0.783 within-domain to 0.563 across domains.
GrACE is a fine-tuned generative method that uses similarity to a special token embedding for real-time calibrated confidence in LLMs and enables efficient confidence-based test-time scaling.
FLaG models hallucination detection via latent evidence groups and energy-based routing with log-marginal aggregation, claiming SOTA results and a theoretical link to Bayes-optimal detection under heterogeneous mechanisms.
Reinforcement learning teaches LLMs to assess their own capabilities more effectively than supervised fine-tuning, preserves original skills, generalizes out of distribution, and aids local-cloud routing and data selection.
Reverse Probing extracts token-level uncertainty from LLM internal activations on labeled clinical summaries, outperforming eight baselines with up to 4x higher AUPRC on two expert-annotated datasets while lowering compute costs.
citing papers explorer
-
Clotho: Measuring Task-Specific Pre-Generation Test Adequacy for LLM Inputs
Clotho ranks LLM test inputs by failure likelihood using pre-generation hidden states and GMMs, achieving 0.716 ROC-AUC after labeling 5.4% of inputs on average across eight tasks and three models, with transfer to proprietary models.