{ "_name_or_path": "GerbilLab/Gerbil-A-32m", "activation_function": "gelu_new", "architectures": [ "GPTJForCausalLM" ], "attn_pdrop": 0.0, "bos_token_id": 50256, "embd_pdrop": 0.0, "eos_token_id": 50256, "initializer_range": 0.02, "layer_norm_epsilon": 1e-05, "model_type": "gptj", "n_embd": 256, "n_head": 8, "n_inner": null, "n_layer": 8, "n_positions": 2048, "resid_pdrop": 0.0, "rotary_dim": 16, "tie_word_embeddings": false, "transformers_version": "4.35.2", "use_cache": false, "vocab_size": 50400 }