DPR-BAG generates factually grounded biomedical abstracts from full texts via structured BOMRC decomposition, parallel LLM prompting, and coherence refinement without any model training.
When Splitting Makes Stronger: A Theoretical and Empirical Analysis of Divide-and-Conquer Prompting in
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Divide-Prompt-Refine: a Training-Free, Structure-Aware Framework for Biomedical Abstract Generation
DPR-BAG generates factually grounded biomedical abstracts from full texts via structured BOMRC decomposition, parallel LLM prompting, and coherence refinement without any model training.