---
dataset_info:
  features:
  - name: layer_7_mlp_down_proj_activations
    sequence: float32
  splits:
  - name: train
    num_bytes: 8390656000
    num_examples: 512000
  download_size: 3329581388
  dataset_size: 8390656000
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---