What are large language models such as GPT designed to do?
Show answer and explanation
Answer: B
Large language models are trained to predict and produce sequences of tokens, which gives them the ability to draft prose, answer questions, summarise, translate and write code from an instruction. Some multimodal variants also generate or interpret images. Their strength is fluent generation grounded in patterns learned from vast corpora, which is why outputs must be reviewed for accuracy. They are not optimised for numeric sorting, query tuning or network monitoring.
Why the other options are wrong
- A. Sorting tables is deterministic data processing and does not require a generative language model.
- C. Query tuning is a database optimisation activity that relies on execution plans, not on text generation.
- D. Network intrusion monitoring is a security analytics and anomaly detection task, not a language generation task.
Exam tip. Prompts in, new content out.