On a CUAD contract QA task, GPT-4.1's exact-match accuracy rises from about 48% to 79% with well-structured Markdown input and a structure-aware system prompt, while GPT-4o stays near 50%.
Mdeval: Evaluating and enhancing markdown awareness in large language models, January 2025
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.CL 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
support 1representative citing papers
citing papers explorer
-
The Hidden Structure -- Improving Legal Document Understanding Through Explicit Text Formatting
On a CUAD contract QA task, GPT-4.1's exact-match accuracy rises from about 48% to 79% with well-structured Markdown input and a structure-aware system prompt, while GPT-4o stays near 50%.