---
dataset_info:
  features:
  - name: embeddings
    sequence: float32
  - name: labels
    sequence: float32
  - name: ground_truth
    dtype: string
  - name: kyutai_text
    dtype: string
  - name: granite_text
    dtype: string
  - name: voxtral_text
    dtype: string
  splits:
  - name: train
    num_bytes: 4927
    num_examples: 2
  download_size: 12173
  dataset_size: 4927
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---