New Dataset

Upload a JSON file containing the corpus of documents you want to analyze.

Required JSON Format

Your file must be a list of objects. Each object must contain an id (string/number) and a text field.

[
  {
    "id": "1",
    "text": "Sparse autoencoders allow us to..."
  },
  {
    "id": "2",
    "text": "Dense embeddings are hard to interpret..."
  }
]