{
  "tokenizer": "declref_symbols",
  "vocab_size": 1028,
  "num_tokens": 9961472,
  "source": "declref",
  "split": "val",
  "dtype": "uint16",
  "gen_params": {
    "seq_length": 2048,
    "num_shards": 256
  }
}