A new evaluation framework with XPath-based labels shows Flat JSON input gives LLMs near-perfect record extraction F1 (0.9567) with low hallucination, far outperforming Slimmed HTML and Hierarchical JSON.
These sections accurately cover the paper’s goal of providing a standard way to evaluate and compare web record extraction systems
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.DB 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
NEXT-EVAL: Next Evaluation of Traditional and LLM Web Data Record Extraction
A new evaluation framework with XPath-based labels shows Flat JSON input gives LLMs near-perfect record extraction F1 (0.9567) with low hallucination, far outperforming Slimmed HTML and Hierarchical JSON.