paper-with-me

Papers

MuDAF: Long-Context Multi-Document Attention Focusing through Contrastive Learning on Attention Heads

2025-02-19 · Weihao Liu, Ning Wu, Shiping Yang, Wenbiao Ding, Shining Liang, Ming Gong, Dongmei Zhang

Large Language Models (LLMs) frequently show distracted attention due to irrelevant information in the input, which severely impairs their long-context capabilities. Inspired by recent studies on the effectiveness of retrieval heads in long-context factutality, we aim at addressing this distraction issue through improving such retrieval heads directly. We propose Multi-Document Attention Focusing (MuDAF), a novel method that explicitly optimizes the attention distribution at the head level through contrastive learning. According to the experimental results, MuDAF can significantly improve the long-context question answering performance of LLMs, especially in multi-document question answering. Extensive evaluations on retrieval scores and attention visualizations show that MuDAF possesses great potential in making attention heads more focused on relevant information and reducing attention distractions.

📄 PDF Abstract BibTeX arXiv:2502.13963

Code (1)

NeosKnight233/MuDAF 공식 구현 jax

Tasks

Contrastive LearningQuestion AnsweringRetrieval

Methods 이 논문이 사용한 방법론

Softmax The Softmax output function transforms a previous layer's output into a vector of probabilities. It is commonly used for multiclass classification. Given an input vector $x$…
Attention 설명 없음

Similar Papers 제목 키워드 기반

AttentionRetriever: Attention Layers are Secretly Long Document Retrievers

2026-02-12 · David Jiahao Fu, Lam Thanh Do, Jiayu Li, Kevin Chen-Chuan Chang arxiv

Retrieval augmented generation (RAG) has been widely adopted to help Large Language Models (LLMs) to process tasks involving long documents. However, existing retrieval models are not designed for long document retrieval…

Attention Sorting Combats Recency Bias In Long Context Language Models

2023-09-28 · Alexander Peysakhovich, Adam Lerer

Current language models often fail to incorporate long contexts efficiently during generation. We show that a major contributor to this issue are attention priors that are likely learned during pre-training: relevant inf…

PositionRetrievalRetrieval-augmented Generation

Understanding Long Documents with Different Position-Aware Attentions

2022-08-17 · Hai Pham, Guoxin Wang, Yijuan Lu, Dinei Florencio 외

Despite several successes in document understanding, the practical task for long document understanding is largely under-explored due to several challenges in computation and how to efficiently absorb long multimodal inp…

document understandingPosition

Leave No Document Behind: Benchmarking Long-Context LLMs with Extended Multi-Doc QA

2024-06-25 · Minzheng Wang, Longze Chen, Cheng Fu, Shengyi Liao 외

Long-context modeling capabilities have garnered widespread attention, leading to the emergence of Large Language Models (LLMs) with ultra-context windows. Meanwhile, benchmarks for evaluating long-context LLMs are gradu…

BenchmarkingLong-Context UnderstandingQuestion AnsweringRAG+1

Attention Expansion: Enhancing Keyphrase Extraction from Long Documents with Attention-Augmented Contextualized Embeddings

2026-06-09 · Roberto Martínez-Cruz, Alvaro J. López-López, José Portela arxiv

Pre-trained language models (PLMs) have achieved strong performance in keyphrase extraction (KPE), largely due to their ability to generate rich contextualized representations. However, long-document KPE remains challeng…

Keyphrase Extraction