t_attention_mask"],
        ...     doc_scores=doc_scores,
        ... )
        >>> generated_string = tokenizer.batch_decode(generated, skip_special_tokens=True)
        ```NFrŠ