SGER applies a two-phase curriculum to fine-tune LLMs for name matching, reporting 99.02% accuracy and 0.994 F1 on 50,000 real-world Indian name pairs while outperforming baselines and deploying in production.
Beyond Full Fine-Tuning: Harnessing the Power of
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
-
Structure-Guided Entity Resolution: Fine-Tuning LLMs for Robust Name Matching in Complex Linguistic Contexts
SGER applies a two-phase curriculum to fine-tune LLMs for name matching, reporting 99.02% accuracy and 0.994 F1 on 50,000 real-world Indian name pairs while outperforming baselines and deploying in production.