Modern generative world models render increasingly realistic action-controllable futures, yet they frequently hallucinate: rollouts remain visually fluent while drifting from the ground-truth dynamics. We hypothesize that hallucination concentrates in low-coverage regions of the state-action space, where lightweight data-centric signals can both detect it and guide mitigation.
arXiv:2606. 27326v1 Announce Type: new Abstract: Modern generative world models render increasingly realistic action-controllable futures, yet they frequently hallucinate: rollouts remain visually fluent while drifting from the ground-truth dynamics.
The paper proposes a hidden‑state probing method for detecting hallucinations at the span level in large language model outputs, moving beyond token‑wise binary classification. By examining layer‑wise activation patterns, the approach identifies the exact onset and continuation tokens of hallucinations, achieving higher precision‑recall AUC than random baselines despite class imbalance. Additionally, the authors introduce a cross‑model detection framework where one model observes another’s internal representations, showing that an external observer can match or surpass the generator’s own self‑detection of hallucination onsets, even when the observer is smaller.
The paper introduces CADMP, a lightweight framework for detecting object hallucinations in large vision‑language models. CADMP measures cross‑modal attention drift between adjacent layers and verifies predictions by masking visually relevant regions, combining these signals to identify hallucinated outputs. Experiments on multiple benchmarks show that CADMP achieves competitive detection performance, and ablation studies confirm the complementary roles of attention drift and mask‑based verification.
Despite recent advances in large vision-language models (LVLMs), object hallucination remains a major barrier to their reliable deployment. Existing detection methods often characterize visual groundi...
arXiv:2602. 01740v3 Announce Type: replace Abstract: Video language models (Video-LLMs) are prone to hallucinations, generating plausible but ungrounded content when visual evidence is weak, ambiguous, or biased.
Large Language Models (LLMs) have demonstrated remarkable capabilities across diverse tasks, yet they remain prone to generating hallucinations. Detecting these hallucinations is critical for deploying LLMs reliably in high-stakes applications.
arXiv:2608. 07302v1 Announce Type: cross Abstract: Large Vision-Language Models (LVLMs) often suffer from object hallucination, generating objects that are absent from the image.
DynHD is a method for detecting hallucinations in diffusion large language models (D‑LLMs) by focusing on token‑level uncertainty and its evolution during the denoising process. It introduces a semantic‑aware evidence construction module that filters out non‑informative structural tokens and highlights uncertainty in informative tokens, and a reference evidence generator that models the expected trajectory of uncertainty, enabling a deviation‑based detector to identify hallucinations. Experiments show DynHD outperforms existing baselines while being more efficient across various benchmarks and backbone models.
The paper introduces a low-cost method for detecting hallucinations in large language models by treating the model as a black-box dynamical system. It projects responses into a high-dimensional manifold, models the latent state-space dynamics with Koopman operator theory, and uses differential residual scores from transition operators to distinguish factual from hallucinated outputs. The approach requires only a single-sample pass and shows state-of-the-art performance across three benchmarks with reduced resource overhead.
arXiv:2606. 24790v1 Announce Type: cross Abstract: Large Language Models (LLMs) have demonstrated remarkable capabilities across diverse tasks, yet they remain prone to generating hallucinations.
arXiv:2608. 10430v1 Announce Type: cross Abstract: Large Language Models (LLMs) deployed as AI agents frequently exhibit user specification-grounding failures, executing hallucinated, undesired actions to force a resolution rather than expressing uncertainty.