Your task is to explain the ANLS* metric for evaluating generative large language models (GLLMs) in document processing tasks. You MUST compare ANLS* to previous evaluation metrics, emphasizing the types of tasks it can evaluate and its benefits. Furthermore, discuss the impact of specialized prompting techniques, such as the SFT method, on the performance of GLLMs. Provide concrete examples to illustrate the advantages of these techniques in improving document processing tasks.