01
Core idea
Large language models are strong generalists. They usually perform better when the task needs complex reasoning, broad world knowledge, flexible instruction following, or careful synthesis across messy context.
Learning nugget
When to use a large model, when a smaller one is enough, and why cost is not the only factor.
Large language models are strong generalists. They usually perform better when the task needs complex reasoning, broad world knowledge, flexible instruction following, or careful synthesis across messy context.