{
  "architectures": [
    "NanoLLMForCausalLM"
  ],
  "d_model": 768,
  "dtype": "float32",
  "eos_token_id": 1,
  "max_seq_len": 256,
  "model_type": "nano_llm",
  "n_heads": 12,
  "n_layers": 6,
  "pad_token_id": 32000,
  "transformers_version": "5.10.2",
  "vocab_size": 32000
}