TriLens: Per-Layer Logit-Lens Entropy for White-Box Hallucination Detection

• URL: http://arxiv.org/abs/2606.01033v1
• Date: Sun, 31 May 2026 05:48:21 GMT
• Title: TriLens: Per-Layer Logit-Lens Entropy for White-Box Hallucination Detection
• Abstract summary: We introduce TriLens, a detector for detecting hallucinations in language models.<n>At every layer, it reads the multi-head self-attention output, the feed-forward output, and the residual stream through the model's own logit lens.<n>The resulting 3L-dimensional trajectory describes how certainty forms across depth and across modules.
• Score: 22.258009855879752
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: When a language model hallucinates, the final answer is wrong, but the mistake is not necessarily invisible inside the model. Different internal pathways may remain uncertain, disagree in how quickly they sharpen, or commit to competing continuations before the output is produced. We introduce TriLens, a white-box detector that turns this intuition into a compact representation: at every layer, it reads the multi-head self-attention output, the feed-forward output, and the residual stream through the model's own logit lens, then records only the entropy of each readout. The resulting 3L-dimensional trajectory describes how certainty forms across depth and across modules, without storing high-dimensional hidden states or sampling multiple generations. This simple signal yields a strong detector across instruction-tuned LLMs and QA benchmarks, and our analyses show that the three module-wise entropy trajectories provide complementary evidence. TriLens suggests that hallucination detection can benefit from tracking how internal computation settles, not only what the final layer predicts.

Related papers

• Bypass Observation: A Conceptual Design of a Non-Intrusive Layer-Wise Semantic Extraction Architecture [2.8943395720203386]
  We introduce Bypass Observation, a non-intrusive layer-wise readout architecture.<n>We derive a closed-form overhead approximation governed primarily by V/(12d)<n>We argue that Bypass Observation can make model iteration more observable while remaining only a partial, potentially misleading projection of hidden states.
  arXiv  Detail & Related papers  (2026-09-12T08:35:45Z)
• HalluTracer: Hallucination Detection via Depth-Averaging Truth Signals [45.64606191137958]
  We introduce HalluTracer, a framework that aggregates truthfulness evidence across every layer of the forward pass.<n>A geometric analysis reveals that the per-layer signals are weakly correlated, so that simple depth averaging suppresses layer-specific noise.<n>HalluTracer consistently outperforms matched white-box baselines.
  arXiv  Detail & Related papers  (2026-08-17T10:03:29Z)
• GhostPoint: Self-Supervised Representation Learning by Hallucinating Occluded LiDAR Structure [58.617405464262646]
  3D object detection from LiDAR point clouds is a core problem in autonomous driving.<n>Recent advances in self-supervised learning (SSL) enable scalable pretraining and transfers well to per-point tasks.<n>We analyze recent SSL methods and find that most objectives are defined only on measured LiDAR returns from visible surfaces.<n>We propose GhostPoint, an SSL framework that hallucinates latent features in local neighborhoods around discovered instances.
  arXiv  Detail & Related papers  (2026-08-14T16:10:35Z)
• UniProbe: A Learnable Token-Level Hallucination Detector for Large VLMs using Multi-Structural Internal Representations [61.85916280854257]
  Large Vision-Language Models (LVLMs) achieve impressive visual reasoning and dialogue capabilities, yet frequently hallucinate content unsupported by the visual input.<n>We introduce textbfUniProbe, a lightweight, unified, learnable detector that models a frozen LVLM's heterogeneous computational trace from a single forward pass.
  arXiv  Detail & Related papers  (2026-08-11T12:01:59Z)
• How Much is Left? LLMs Linearly Encode Their Remaining Output Length [3.541590945326217]
  Large language models generate one token at a time, yet their responses show remarkably consistent length structure.<n>We ask whether the model carries an internal estimate of how much response remains.<n>We train minimal-capacity linear probes on frozen hidden states of three open-weight 7-8B models across seven completion-style datasets.
  arXiv  Detail & Related papers  (2026-07-06T16:56:04Z)
• EnsemHalDet: Robust VLM Hallucination Detection via Ensemble of Internal State Detectors [1.4401311275746886]
  Vision-Language Models (VLMs) excel at multimodal tasks, but they remain vulnerable to hallucinations that are factually incorrect or ungrounded in the input image.<n>Recent work suggests that hallucination detection using internal representations is more efficient and accurate than approaches that rely solely on model outputs.<n>We propose EnsemHalDet, an ensemble-based hallucination detection framework.
  arXiv  Detail & Related papers  (2026-04-03T06:48:27Z)
• DRIFT: Detecting Representational Inconsistencies for Factual Truthfulness [5.785021425715989]
  LLMs often produce fluent but incorrect answers, yet detecting such hallucinations typically requires multiple sampling passes or post-hoc verification.<n>We propose a lightweight probe to read these signals directly from hidden states.<n>We develop an LLM router that answers confident queries immediately while delegating uncertain ones to stronger models.
  arXiv  Detail & Related papers  (2026-01-20T18:16:10Z)
• EigenTrack: Spectral Activation Feature Tracking for Hallucination and Out-of-Distribution Detection in LLMs and VLMs [8.616813040714883]
  EigenTrack is an interpretable real-time detector for large language models (LLMs)<n>It tracks temporal shifts in representation structure that signal hallucination and OOD drift before surface errors appear.<n>Unlike existing white-box detectors, it preserves temporal context, aggregates global signals, and offers interpretable accuracy-latency trade-offs.
  arXiv  Detail & Related papers  (2025-09-19T08:05:28Z)
• HalluLens: LLM Hallucination Benchmark [49.170128733508335]
  Large language models (LLMs) often generate responses that deviate from user input or training data, a phenomenon known as "hallucination"<n>This paper introduces a comprehensive hallucination benchmark, incorporating both new extrinsic and existing intrinsic evaluation tasks.
  arXiv  Detail & Related papers  (2025-04-24T13:40:27Z)
• Generate, but Verify: Reducing Hallucination in Vision-Language Models with Retrospective Resampling [78.78822033285938]
  Vision-Language Models (VLMs) excel at visual understanding but often suffer from visual hallucinations.<n>In this work, we introduce REVERSE, a unified framework that integrates hallucination-aware training with on-the-fly self-verification.
  arXiv  Detail & Related papers  (2025-04-17T17:59:22Z)
• Logical Closed Loop: Uncovering Object Hallucinations in Large Vision-Language Models [52.957842999317506]
  Object hallucination refers to the phenomenon that the LVLMs claim non-existent objects in the image. We propose a Logical Closed Loop-based framework for Object Hallucination Detection and Mitigation, namely LogicCheckGPT. As a plug-and-play method, it can be seamlessly applied to all existing LVLMs.
  arXiv  Detail & Related papers  (2024-02-18T15:28:39Z)
• SelfCheckGPT: Zero-Resource Black-Box Hallucination Detection for Generative Large Language Models [55.60306377044225]
  "SelfCheckGPT" is a simple sampling-based approach to fact-check the responses of black-box models. We investigate this approach by using GPT-3 to generate passages about individuals from the WikiBio dataset.
  arXiv  Detail & Related papers  (2023-03-15T19:31:21Z)
• Don't Say What You Don't Know: Improving the Consistency of Abstractive Summarization by Constraining Beam Search [54.286450484332505]
  We analyze the connection between hallucinations and training data, and find evidence that models hallucinate because they train on target summaries that are unsupported by the source. We present PINOCCHIO, a new decoding method that improves the consistency of a transformer-based abstractive summarizer by constraining beam search to avoid hallucinations.
  arXiv  Detail & Related papers  (2022-03-16T07:13:52Z)

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.