A small language model (SLM) fine-tuned by Distil Labs for policy-aware PII redaction that outputs a single JSON object with redactedtext and entities`. Optimized to run locally with strong accuracy and strict schema adherence. Developed by: Distil Labs GmbH License: Llama 3.2 Community License Agreement Finetuned from: meta-llama/Llama-3.2-3B-Instruct` Use cases: Redacting support chats, logs, tickets, transcripts—removing identity while preserving ops signals (IDs last-4, order numbers, etc.). Out of scope: Legal or compliance advice; languages beyond English (generalization not guaranteed); domain-specific IDs unseen in training. Input: A plain-text prompt with task instruction + context. INVALID LANGUAGE PAIR SPECIFIED. EXAMPLE: LANGPAIR=EN|IT USING 2 LETTER ISO OR RFC3066 LIKE ZH-CN. ALMOST ALL LANGUAGES SUPPORTED BUT SOME MAY HAVE NO CONTENT Judged by a frontier LLM using a deterministic rubric: JSON-only, schema validity, redactedtext exact match, and set-equality of (value, replacementtoken) pairs (reason/order ignored). Score: 0.82…
Модальности:
Генерация текста
Области применения:
Диалог / чат Следование инструкциям
Задача: Генерация текста
Автор: distil-labs
Теги: llama, pii-redaction, privacy, slm, distil-labs, conversational, en
Лайков: 3 | Загрузок: 44
Описание основано на материалах HuggingFace. Перевод выполнен автоматически.
