Detecting Contextual Hallucinations in LLMs with Frequency-Aware Attention
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866917284083138560 |
|---|---|
| author | Qi, Siya Chen, Yudong Zhao, Runcong Zhu, Qinglin Hu, Zhanghao Liu, Wei He, Yulan Yuan, Zheng Gui, Lin |
| author_facet | Qi, Siya Chen, Yudong Zhao, Runcong Zhu, Qinglin Hu, Zhanghao Liu, Wei He, Yulan Yuan, Zheng Gui, Lin |
| contents | Hallucination detection is critical for ensuring the reliability of large language models (LLMs) in context-based generation. Prior work has explored intrinsic signals available during generation, among which attention offers a direct view of grounding behavior. However, existing approaches typically rely on coarse summaries that fail to capture fine-grained instabilities in attention. Inspired by signal processing, we introduce a frequency-aware perspective on attention by analyzing its variation during generation. We model attention distributions as discrete signals and extract high-frequency components that reflect rapid local changes in attention. Our analysis reveals that hallucinated tokens are associated with high-frequency attention energy, reflecting fragmented and unstable grounding behavior. Based on this insight, we develop a lightweight hallucination detector using high-frequency attention features. Experiments on the RAGTruth and HalluRAG benchmarks show that our approach achieves performance gains over verification-based, internal-representation-based, and attention-based methods across models and tasks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_18145 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Detecting Contextual Hallucinations in LLMs with Frequency-Aware Attention Qi, Siya Chen, Yudong Zhao, Runcong Zhu, Qinglin Hu, Zhanghao Liu, Wei He, Yulan Yuan, Zheng Gui, Lin Computation and Language Hallucination detection is critical for ensuring the reliability of large language models (LLMs) in context-based generation. Prior work has explored intrinsic signals available during generation, among which attention offers a direct view of grounding behavior. However, existing approaches typically rely on coarse summaries that fail to capture fine-grained instabilities in attention. Inspired by signal processing, we introduce a frequency-aware perspective on attention by analyzing its variation during generation. We model attention distributions as discrete signals and extract high-frequency components that reflect rapid local changes in attention. Our analysis reveals that hallucinated tokens are associated with high-frequency attention energy, reflecting fragmented and unstable grounding behavior. Based on this insight, we develop a lightweight hallucination detector using high-frequency attention features. Experiments on the RAGTruth and HalluRAG benchmarks show that our approach achieves performance gains over verification-based, internal-representation-based, and attention-based methods across models and tasks. |
| title | Detecting Contextual Hallucinations in LLMs with Frequency-Aware Attention |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2602.18145 |