tor (`float`, *optional*, defaults to 0.002):
            A factor for initializing all weight matrices.
        output_router_logits (`bool`, *optional*, default to `False`):
            Whether or not to return the router logits of all experts.
        use_cache (`bool`, *optional*, defaults to `True`):
            Whether or not the model should return the last key/values attentions (not used by all models)
    zgptsan-japaneseZpast_key_valuesÚ