paper-with-me

Papers

Retrieval Heads are Dynamic

2026-01-07 · Yuping Lin, Zitao Li, Yue Xing, Pengfei He, Yingqian Cui, Yaliang Li, Bolin Ding, Jingren Zhou, Jiliang Tang arxiv

Recent studies have identified "retrieval heads" in Large Language Models (LLMs) responsible for extracting information from input contexts. However, prior works largely rely on static statistics aggregated across datasets, identifying heads that perform retrieval on average. This perspective overlooks the fine-grained temporal dynamics of autoregressive generation. In this paper, we investigate retrieval heads from a dynamic perspective. Through extensive analysis, we establish three core claims: (1) Dynamism: Retrieval heads vary dynamically across timesteps; (2) Irreplaceability: Dynamic retrieval heads are specific at each timestep and cannot be effectively replaced by static retrieval heads; and (3) Correlation: The model's hidden state encodes a predictive signal for future retrieval head patterns, indicating an internal planning mechanism. We validate these findings on the Needle-in-a-Haystack task and a multi-hop QA task, and quantify the differences on the utility of dynamic and static retrieval heads in a Dynamic Retrieval-Augmented Generation framework. Our study provides new insights into the internal mechanisms of LLMs.

📄 PDF Abstract BibTeX arXiv:2602.11162

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Can Retrieval Heads See Images? Multimodal Retrieval Heads in Long-Context Vision-Language Models

2026-05-26 · Aaron Branson Cigres Li, Zhaowei Wang, Yu Zhao, Yiming Du 외 arxiv

Large vision-language models increasingly rely on long-context modeling to reason over documents, hour-level videos, and long-horizon agent trajectories, requiring them to locate relevant evidence across interleaved text…

Image RetrievalHead DetectionText Retrieval

Retrieval Head Mechanistically Explains Long-Context Factuality

2024-04-24 · Wenhao Wu, Yizhong Wang, Guangxuan Xiao, Hao Peng 외

Despite the recent progress in long-context language models, it remains elusive how transformer-based models exhibit the capability to retrieve relevant information from arbitrary locations within the long context. This …

Continual PretrainingHallucinationLarge Language ModelLong-Context Understanding+1

HeteroCache: A Dynamic Retrieval Approach to Heterogeneous KV Cache Compression for Long-Context LLM Inference

2026-01-20 · Zhiyuan Shi, Qibo Qiu, Feng Xue, Zhonglin Jiang 외 arxiv

The linear memory growth of the KV cache poses a significant bottleneck for LLM inference in long-context tasks. Existing static compression methods often fail to preserve globally important information. Although recent …

DeCoRe: Decoding by Contrasting Retrieval Heads to Mitigate Hallucinations

2024-10-24 · Aryo Pradipta Gema, Chen Jin, Ahmed Abdulaal, Tom Diethe 외

Large Language Models (LLMs) often hallucinate, producing unfaithful or factually incorrect outputs by misrepresenting the provided context or incorrectly recalling internal knowledge. Recent studies have identified spec…

Instruction FollowingQuestion AnsweringRetrieval

VERA: Identifying and Leveraging Visual Evidence Retrieval Heads in Long-Context Understanding

2026-02-09 · Rongcan Pei, Huan Li, Fang Guo, Qi Zhu arxiv

While Vision-Language Models (VLMs) have shown promise in textual understanding, they face significant challenges when handling long context and complex reasoning tasks. In this paper, we dissect the internal mechanisms …

Long-Context Understanding