README.md

metadata

language:
  - en
license: mit
dataset_info:
  features:
    - name: source
      dtype: string
    - name: author
      dtype: string
    - name: title
      dtype: string
    - name: description
      dtype: string
    - name: url
      dtype: string
    - name: urlToImage
      dtype: string
    - name: publishedAt
      dtype: string
    - name: content
      dtype: string
    - name: category_nist
      dtype: string
    - name: category
      dtype: string
    - name: id
      dtype: string
    - name: subreddit
      dtype: string
    - name: score
      dtype: int64
    - name: num_comments
      dtype: int64
    - name: created_time
      dtype: timestamp[ns]
    - name: top_comments
      dtype: string
  splits:
    - name: train
      num_bytes: 649243675
      num_examples: 93259
  download_size: 364163308
  dataset_size: 649243675
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*

REALM: REAL-World Application of Large Language Models

Dataset Description

• Paper: coming soon

• Dashboard Demo: https://realm-e7682.web.app/

• License: [mit]

• Language(s) (NLP): English

• Point of Contact: Jingwen

Dataset Summary

Large Language Models (LLMs), such as GPT-like models, have transformed industries and everyday life, creating significant societal impact. To better understand their real-world applications, we created the REALM Dataset, a collection of over 93k use cases sourced from Reddit posts and news articles, spanning 2020-06(when GPT was first released) to 2024-12. REALM focuses on two key aspects:

1. How LLMs are being used: Categorizing the wide range of applications, following AI Use Taxonomy: A Human-Centered Approach.

2. Who is using them: Extracting the occupation attributes of current or potential end-users, categorized based on the O*NET classification system.

Updates

2025-2-15: Content Update. Paper submitted to ACL 2025.

Languages

English

Data Fields

• `` (string):

Citation Information

Please consider citing our paper if you find this dataset useful:

@inproceedings{
  \\\
}