sae_lens.json

{
  "sae_lens_version": "2.1.2",
  "tokenizer_name": "gpt2",
  "original_dataset": "NeelNanda/c4-10k",
  "original_split": "train",
  "original_data_files": null,
  "context_size": 128,
  "shuffled": true,
  "seed": null,
  "begin_batch_token": "bos",
  "begin_sequence_token": null,
  "sequence_separator_token": "eos"
}