Instruction-tuned open-source LLMs, especially DeepSeek-Q1, outperform fine-tuned, RAG, and NER baselines on PII redaction accuracy and leakage in this benchmark.
Natural Language Processing – Amazon Comprehend,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CR 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
PRvL: Quantifying the Capabilities and Risks of Large Language Models for PII Redaction
Instruction-tuned open-source LLMs, especially DeepSeek-Q1, outperform fine-tuned, RAG, and NER baselines on PII redaction accuracy and leakage in this benchmark.