Metraipolis › archive › dc_78810f3719cf
Do LLMs Understand Context? A Knowledge Graph-Based Evaluation Framework — arXiv:2609.30484
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.30484
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-08T01:38:10.284Z
- SHA-256 of the archived text
9dc2494b36d2495ad4e62b211411e2a55b0e583b0c42087640006f03becaa90a- Stable URL of this page
https://metraipolis.com/documento/dc_78810f3719cf
How to cite
Do LLMs Understand Context? A Knowledge Graph-Based Evaluation Framework — arXiv:2609.30484 — arXiv (en), https://arxiv.org/abs/2609.30484 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_78810f3719cf (Metraipolis archive dc_78810f3719cf, sha256 9dc2494b36d2495ad4e62b211411e2a55b0e583b0c42087640006f03becaa90a)
BibTeX
@misc{dc78810f3719cf,
title = {Do LLMs Understand Context? A Knowledge Graph-Based Evaluation Framework — arXiv:2609.30484},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.30484},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_78810f3719cf; sha256 9dc2494b36d2495ad4e62b211411e2a55b0e583b0c42087640006f03becaa90a},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.30484. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
While large language models (LLMs) have achieved remarkable linguistic capabilities, a profound question lingers at their core: do these models truly comprehend context or simply excel at pattern matching on an unprecedented scale? Contextual understanding in LLMs refers to the ability to correctly extract relevant information from a given context, integrate it into a coherent internal representation, and reason over it to produce factually consistent and contextually grounded responses. However, traditional methods such as BiLingual Evaluation Understudy (BLEU) and perplexity simply measure surface-level performance. This reveals a critical gap in question answering (QA), where responses must be contextually grounded rather than simply being memorized associations. To fill this void, we propose a novel knowledge graph (KG) based evaluation framework for LLM contextual understanding in QA. Central to this is Semantic Structural Similarity for KGs (S3KG), a hybrid similarity measure combining structural and semantic signals into a single score. In addition, a diagnostic analysis framework is developed to identify and categorize reasoning errors at the triplet level, enabling fine-grained analysis of model failures. Together, across nine benchmarks, S3KG achieves F1 gains of up to $+7.6$ points over the strongest baseline and AUROC up to $0.973$. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-28 in cs.AI). Categorie arXiv: Artificial Intelligence (cs.AI) ; Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2609.30484 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-08T01:12:14.242Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-28?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).