CREFT, a sequential multi-agent LLM pipeline, improves character relation extraction over a single-agent baseline on an internal Korean drama dataset, though the gains lack statistical significance and the dataset is not public.
If any specific relationships occur too frequently and disrupt diversity in the relationships list, select these for review
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
CREFT: Sequential Multi-Agent LLM for Character Relation Extraction
CREFT, a sequential multi-agent LLM pipeline, improves character relation extraction over a single-agent baseline on an internal Korean drama dataset, though the gains lack statistical significance and the dataset is not public.