問題文
A test team wants an LLM to summarize a long piece of product documentation in a single request. Someone proposes moving to a model with a much larger context window. Which statement matches what the syllabus says about that choice?
選択肢
- The model can hold the coherence across the longer documentation passages, but the added tokens raise the computational effort and the response time.
- The model can learn the terminology across the longer documentation passages, but the added tokens raise the training expense and the labeling cost.
- The model can widen the vocabulary across the longer documentation passages, but the added tokens raise the licensing charge and the storage need.
- The model can skip the tokenization across the longer documentation passages, but the added tokens raise the annotation effort and the review load.