Atomic propositions improve relation recall for weak triplet extractors like GLiREL and small models across SMiLER, FewRel, DocRED and CaRB, while requiring fallback for stronger LLMs.
As a baseline, we replace stage 2 with two sub- stages by using the Parsing and Triplet Extrac- tion as follows: 2.1
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
LLM-based Atomic Propositions help weak extractors: Evaluation of a Propositioner for triplet extraction
Atomic propositions improve relation recall for weak triplet extractors like GLiREL and small models across SMiLER, FewRel, DocRED and CaRB, while requiring fallback for stronger LLMs.