DMRL applies inverse reinforcement learning with GRPO to train LLMs to extract PII, reporting higher reconstruction accuracy than four baselines on three datasets.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
DMRL: Data- and Model-aware Reward Learning for Data Extraction
DMRL applies inverse reinforcement learning with GRPO to train LLMs to extract PII, reporting higher reconstruction accuracy than four baselines on three datasets.