AI and LLMs · Middle · Question 46 of 75
Why count tokens before sending a request?
- 1To stay within context limits and predict cost
- 2To make the model's answer more accurate
- 3Because providers reject requests with odd token counts
Show the answer
Correct answer: To stay within context limits and predict cost
Explanation
Libraries or the provider's token counting endpoint let apps trim history or documents before hitting limits.