

















Abstract:Multimodal large language models (MLLMs) have achieved remarkable progress, yet the object hallucination remains a critical challenge for reliable deployment. In this paper, we present an in-depth analysis of instruction token embeddings and reveal that they implicitly encode visual information while effectively filtering erroneous information introduced by misleading visual embeddings. Building on this insight, we propose the Instruction Lens Score (InsLen), which combines a Calibrated Local Score with a Context Consistency Score that measures context consistency of the object tokens. The proposed approach serves as a plug-and-play object hallucination detector without relying on auxiliary models or additional training. Extensive experiments across multiple benchmarks and diverse MLLM architectures demonstrate that InsLen consistently outperforms existing hallucination detection methods, highlighting its effectiveness and robustness. The code is available at this https URL.
| Comments: | Accepted by ICML-2026 |
| Subjects: | Machine Learning (cs.LG) |
| Cite as: | arXiv:2605.12258 [cs.LG] |
| (or arXiv:2605.12258v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2605.12258 arXiv-issued DOI via DataCite (pending registration) |
From: Runhe Lai [view email]
[v1]
Tue, 12 May 2026 15:27:40 UTC (2,734 KB)
此内容由惯性聚合(RSS阅读器)自动聚合整理,仅供阅读参考。 原文来自 — 版权归原作者所有。