Lune

ACL Findings2025

HG-InsightLog: Context Prioritization and Reduction for Question Answering with Non-Natural Language Construct Log Data

Supriya Bajpai, Athira Gopal, Chandrakant Harjpal, Niraj Kumar

Abstract

Modern IT systems generate vast amounts of log data, which pose challenges for Large Language Models (LLMs) due to their large size, irrelevant entries, and non-Natural Language (non-NL) construct (e.g., domain-specific jargon, error codes, file paths, and abbreviations). Traditional methods like Retrieval-Augmented Generation (RAG) and GraphRAG fail to preserve temporal sequences, handle non-NL for context and entities extraction, and dynamically prioritize query-relevant context. To address these limitations, we propose HG-InsightLog, a novel framework that constructs a multi-entity temporal hypergraph representing log attributevalue pair as nodes and connecting them with hyperedges, capturing critical connections in the data. HG-InsightLog introduces a multistep query personalization mechanism enhancing the Personalized PageRank algorithm to rank hyperedges based on query relevance and contextual centrality to priortize critical connections. Top ranked hyperedges are extracted and converted back into log formats preserving temporal order and reducing context. Experimental results across multiple datasets demonstrate its superiority over existing methods, enhancing factual, causal, and analytical reasoning. Our approach enables smaller LLMs like LLaMA-8B to perform effective log-based QA. Being model-agnostic and training-free, it scales with evolving open-source LLMs without relying on proprietary systems.

Ask Lune

Know this paper in minutes, not an afternoon.

Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.

Questions to start from

Lune can draw on

the full paper · related work

Get the answer

Free to start. No credit card required.

Lineage

  • HippoRAG: Neurobiologically Inspired Long-Term Memory for Large Language Models cites Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks
Builds onCited byLeft to right by year; the more cited, the higher.

Builds on9

  1. Retrieval-Augmented Generation for Knowledge-Intensive NLP TasksPatrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni et al.NeurIPS 2020 · 19,162 citations
  2. BERTScore: Evaluating Text Generation with BERTTianyi Zhang, Varsha Kishore, Felix Wu, Kilian Q. Weinberger et al.ICLR 2020 · 8,443 citations
  3. HippoRAG: Neurobiologically Inspired Long-Term Memory for Large Language ModelsBernal Jimenez Gutierrez, Yiheng Shu, Yu Gu, Michihiro Yasunaga et al.NeurIPS 2024 · 395 citations
  4. LightRAG: Simple and Fast Retrieval-Augmented GenerationZirui Guo, Lianghao Xia, Yanhua Yu, Tu Ao et al.EMNLP Findings 2025 · 209 citations
  5. Automatic Root Cause Analysis via Large Language Models for Cloud IncidentsYinfang Chen, Huaibing Xie, Minghua Ma, Yu Kang et al.EuroSys 2024 · 175 citations

Related papers

  1. You Don't Need Pre-Built Graphs for RAG: Retrieval Augmented Generation with Adaptive Reasoning StructuresShengyuan Chen, Chuang Zhou, Zheng Yuan, Qinggang Zhang et al.AAAI 2026 · 14 citations
  2. Query-Aware Knowledge Retrieval via Hyperbolic StructuringChuang Zhou, Junnan Dong, Yilin Xiao, Shengyuan Chen et al.ACL 2026
  3. HyperG: Hypergraph-Enhanced LLMs for Structured KnowledgeSirui Huang, Hanqian Li, Yanggan Gu, Xuming Hu et al.SIGIR 2025 · 4 citations
  4. OG-RAG: Ontology-grounded retrieval-augmented generation for large language modelsKartik Sharma, Peeyush Kumar, Yunqing LiEMNLP 2025 · 6 citations
  5. Empowering GraphRAG with Knowledge Filtering and IntegrationKai Guo, Harry Shomer, Shenglai Zeng, Haoyu Han et al.EMNLP 2025 · 2 citations
Dusk over the sea between two cliffs drawn in fine vertical lines