Free AWS AIF-C01 Certified AI Practitioner Actual Exam Questions - Question 4 Discussion

Option A—tokens are the actual text chunks, not their numeric forms.

I'd toss out C and D right away since tokens aren't weights or prompts. Between A and B, tokens are definitely the text chunks before any math conversion, so B feels like a step later in the pipeline. Does that make sense?

A/B? Tokens aren’t the embeddings themselves but they’re more than just words—they can be subwords or pieces. So A covers the real text units, while B mixes in the math part, which feels off for this question.

Maybe A, since tokens are the pieces AI breaks text into before processing.

Maybe A here. Tokens are definitely the chunks AI reads before turning into embeddings, so they’re not the numerical vectors themselves like B suggests.

A/B? A makes sense since tokens include subwords, not just full words. B talks about mathematical representations, but that’s more like embeddings, not tokens themselves.

A sounds right but B kinda fits too?