lilacai/mosaicml-local-mosaic-instruct-v3

This dataset is generated by [Lilac](http://lilacml.com) for a HuggingFace Space: [huggingface.co/spaces/lilacai/mosaicml](https://huggingface.co/spaces/lilacai/mosaicml). Original dataset: [https://huggingface.co/datasets/mosaicml/instruct-v3](https://huggingface.co/datasets/mosaicml/instruct-v3) Lilac dataset config: ```embeddings: - {embedding: gte-small, path: response} - {embedding: gte-small, path: prompt} name: mosaic-instruct-v3 namespace: local settings: preferred_embedding: gte-small ui: media_paths: [prompt, response] signals: - path: prompt signal: {signal_name: pii} - path: prompt signal: {signal_name: text_statistics} - path: prompt signal: {signal_name: near_dup} - path: prompt signal: {signal_name: lang_detection} - path: prompt signal: {concept_name: negative-sentiment, embedding: gte-small, namespace: lilac, signal_name: concept_score} - path: prompt signal: {concept_name: non-english, embedding: gte-small, namespace: lilac, signal_name: concept_score} - path: prompt signal: {concept_name: profanity, embedding: gte-small, namespace: lilac, signal_name: concept_score} - path: prompt signal: {concept_name: source-code, embedding: gte-small, namespace: lilac, signal_name: concept_score} - path: prompt signal: {concept_name: toxicity, embedding: gte-small, namespace: lilac, signal_name: concept_score} - path: response signal: {signal_name: pii} - path: response signal: {signal_name: text_statistics} - path: response signal: {signal_name: near_dup} - path: response signal: {signal_name: lang_detection} - path: response signal: {concept_name: negative-sentiment, embedding: gte-small, namespace: lilac, signal_name: concept_score} - path: response signal: {concept_name: non-english, embedding: gte-small, namespace: lilac, signal_name: concept_score} - path: response signal: {concept_name: profanity, embedding: gte-small, namespace: lilac, signal_name: concept_score} - path: response signal: {concept_name: source-code, embedding: gte-small, namespace: lilac, signal_name: concept_score} - path: response signal: {concept_name: toxicity, embedding: gte-small, namespace: lilac, signal_name: concept_score} source: {dataset_name: mosaicml/instruct-v3, source_name: huggingface} ```

原始信息汇总

数据集概述

数据集配置

• 嵌入配置:
  • 嵌入类型: gte-small
  • 路径: response 和 prompt
• 名称: mosaic-instruct-v3
• 命名空间: local
• 设置:
  • 首选嵌入: gte-small
  • UI媒体路径: [prompt, response]

信号配置

• 路径: prompt

  • 信号: pii
  • 信号: text_statistics
  • 信号: near_dup
  • 信号: lang_detection
  • 概念信号: negative-sentiment
  • 概念信号: non-english
  • 概念信号: profanity
  • 概念信号: source-code
  • 概念信号: toxicity

• 路径: response

  • 信号: pii
  • 信号: text_statistics
  • 信号: near_dup
  • 信号: lang_detection
  • 概念信号: negative-sentiment
  • 概念信号: non-english
  • 概念信号: profanity
  • 概念信号: source-code
  • 概念信号: toxicity

数据源

• 数据集名称: mosaicml/instruct-v3
• 来源名称: huggingface