KG-MuLQA: A Framework for KG-based Multi-Level QA Extraction and Long-Context LLM Evaluation

Nikita Tatarinov, Vidhyakshaya Kannan, Haricharana Srinivasa, Arnav Raj, Harpreet Singh Anand +5 more
5/18/2025
cs.CL

No official code found. Best implementation: gtfintechlab/KG-MuLQA (2 stars, last commit 2026-01-13).

Abstract

We introduce KG-MuLQA (Knowledge-Graph-based Multi-Level Question-Answer Extraction): a framework that (1) extracts QA pairs at multiple complexity levels (2) along three key dimensions -- multi-hop retrieval, set operations, and answer plurality, (3) by leveraging knowledge-graph-based document representations. This approach enables fine-grained assessment of model performance across controlled difficulty levels. Using this framework, we construct a dataset of 20,139 QA pairs based on financial credit agreements and evaluate 16 proprietary and open-weight Large Language Models, observing that even the best-performing models struggle with set-based comparisons and multi-hop reasoning over long contexts. Our analysis reveals systematic failure modes tied to semantic misinterpretation and inability to handle implicit relations.

View on arXivView PDF

Code Implementations(1)

gtfintechlab/KG-MuLQACommunity100%
21HTML, PythonMay 15, 20259 months agoNOASSERTION

Similar Papers

The benefits of query-based KGQA systems for complex and temporal questions in LLM era

87% match1 repo22025

KG-Rank: Enhancing Large Language Models for Medical QA with Knowledge Graphs and Ranking Techniques

85% match1 repo502024

A Human-in-the-Loop, LLM-Centered Architecture for Knowledge-Graph Question Answering

84% match2026

Large Language Models Meet Knowledge Graphs for Question Answering: Synthesis and Opportunities

84% match1 repo1632025

Diagnosing and Addressing Pitfalls in KG-RAG Datasets: Toward More Reliable Benchmarking

84% match2025

Ground-Truth Subgraphs for Better Training and Evaluation of Knowledge Graph Augmented LLMs

83% match2025

Cite this paper

@article{tatarinov2025kgmulqa,
  title  = {KG-MuLQA: A Framework for KG-based Multi-Level QA Extraction and Long-Context LLM Evaluation},
  author = {Nikita Tatarinov and Vidhyakshaya Kannan and Haricharana Srinivasa and Arnav Raj and Harpreet Singh Anand and Varun Singh and Aditya Luthra and Ravij Lade and Agam Shah and Sudheer Chava},
  year   = {2025},
  eprint = {2505.12495},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2505.12495v2}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2505.12495.svg)](https://codeofpaper.com/paper/2505.12495)

Discussion