Lost-in-the-Middle Effect

Latest papers

11 papers1 paper in the last four weeks, with none the four weeks before. 0.0% of all new papers.

Query-Guided RoPE Modulation for Lost-in-the-Middle Mitigation in Long-Context LLMs
Nov 6, 2026Peiyu Zhang, Heng Ping, Shinibali Bhattacharyya +6Lost-in-the-Middle Effect Long-Context Language Model Inference
University of Southern California · Amazon Web Services
DuQUAD: A Dual-View Framework with Quality-Aware Evidence Pruning for Multi-Document Question Answering
Oct 25, 2026Yeonjun Choi, Seyeon Jeong, Jong Wook Kim +1Multi-Document QA Evidence-Grounded QA
Graduate School of Information, Yonsei University, Seoul, South Korea · Department of Computer Science, Sangmyung University, Seoul, South Korea
An Adjoint-Sensitivity Framework for Lost-in-the-Middle Phenomena in Causal Residual Transformers
Jul 20, 2026Cheng Huan, Hongwei YuanLost-in-the-Middle Effect Transformer Attention
Department of Statistics and Data Science, The Chinese University of Hong Kong. · Department of Mathematics, University of Macau.
LongCrafter: Towards Diverse Long-Context Understanding via Evidence-Graph-Guided Instruction Synthesis
Jul 7, 2026Chenhao Yuan, Yinhao Xu, Shuwen Xu +8Lost-in-the-Middle Effect LLM Fine-Tuning
University of Chinese Academy of Sciences, Beijing, China · The Key Laboratory of Cognition and Decision Intelligence for Complex Systems, Institute of Automation, Chinese Academy of Sciences, Beijing, China
Prefilling-dLLM: Predictive Prefilling for Long-Context Inference in Diffusion Language Models
Jun 9, 2026Jing Xiong, Qi Han, Shansan Gong +5Lost-in-the-Middle Effect Prefill-Decode Disaggregation
The University of Hong Kong · University of Michigan, Ann Arbor · LMSYS Org
Kinetic theory for Transformers and the lost-in-the-middle phenomenon
May 9, 2026Mitia Duerinckx, Borjan Geshkovski, Stefano RossiLost-in-the-Middle Effect Self-Attention
BERAG: Bayesian Ensemble Retrieval-Augmented Generation for Knowledge-based Visual Question Answering
Apr 24, 2026Jinghong Chen, Jingbiao Mei, Guangyu Yang +1Visual Question Answering Retrieval-Augmented Generation
Department of Engineering, University of Cambridge
Escaping the Context Bottleneck: Active Context Curation for LLM Agents via Reinforcement Learning
Apr 13, 2026Xiaozhe Li, Tianyi Lyu, Yizhao Yang +6Lost-in-the-Middle Effect LLM Context Management
Tongji University · CurrentsAI Research · Stanford University +1
Self-Describing Structured Data with Dual-Layer Guidance: A Lightweight Alternative to RAG for Precision Retrieval in Large-Scale LLM Knowledge Navigation
Mar 28, 2026Hung Ming LiuLost-in-the-Middle Effect Information Retrieval
PARRAWA AI
Lost in the Middle at Birth: An Exact Theory of Transformer Position Bias
Mar 10, 2026Borun D ChowdhuryLost-in-the-Middle Effect Transformer
Meta, London, W1T 1FB, UK
A Structural Theory of Position Bias in Transformers
Feb 18, 2026Hanna Herasimchyk, Robin Labryga, Tomislav Prusina +1Lost-in-the-Middle Effect Position Bias
University of Hamburg Hamburg, Germany

Nov 6, 2026J/K move · Enter open · S save

Query-Guided RoPE Modulation for Lost-in-the-Middle Mitigation in Long-Context LLMs

Peiyu Zhang, Heng Ping, Shinibali Bhattacharyya, Anwesan Pal+5

University of Southern California · Amazon Web Services

Lost-in-the-Middle Effect Long-Context Language Model Inference Positional Encoding Rotary Positional Embeddings Long-Context Retrieval

Open paper AskPDF arXiv

Long-context language models often neglect information in the middle of the input, a phenomenon known as "lost-in-the-middle". Part of this gap may stem from the long-range decay of rotary position embedding (RoPE), which weakens the positional signal between distant tokens. Existing input compression deletes low-importance tokens, and attention reweighting acts only after RoPE has weakened this signal. RoPE-level methods modify positions directly but apply the same remapping to every query. To address these limitations, we propose QGPM (Query-Guided Positional Modulation), a training-free method that reshapes RoPE positions using query relevance from the model's own attention heads. QGPM builds a broad relevance map for lower layers and a sharp one for upper layers. Low-relevance tokens advance by less than one position on fast-rotating planes, pulling relevant evidence toward the query. Slow-rotating planes keep the original layout, with every token in place. Experiments on NoLiMa and RULER with multiple context lengths show that QGPM improves over base models and strong baselines, with the largest gains on retrieval tasks, and generalizes across multiple backbones. Our code is available at https://github.com/iZHpy/QGPM.