Model-agnostic PII detection with LLMs
TL;DR
A configurable, model-agnostic detector that turns any large language model on Amazon Bedrock into a PII detector. Because the entities to detect live in a prompt rather than in code, one detector adapts to new entity types without retraining, and it outperforms an off-the-shelf tool across five public corpora and nine LLM-based detectors.
Nauti's Take
Configuring a PII detector through the prompt instead of code is a practical win, since new entity types arrive without retraining and the underlying model stays swappable. The catch is that detection quality now rides on prompt wording, and with sensitive data a handful of misses gets expensive fast.
Teams with shifting compliance needs gain flexibility, while regulated core processes still need hard evaluation.
Sources
