Round 12: Tossup 18

The number of these things is the numerator of a ratio that is empirically optimal at about 20, the Chinchilla point. Computations involving the Q, K, and V vectors generated from these things are stored in a KV cache, (-5[1])which improves a metric called the “time to the first” one of these things. (-5[1])The process of converting an input into these (10[1])things is commonly done using byte-pair encoding. (10[1])During the prefill stage, IDs corresponding to these things are embedded into a vector that is then passed through multiple transformer layers. The number of these things that can be processed at once is the context window. (10[2])The cost of AI compute (10[1])is often reported (10[1])as the price per one million of these things. (-5[1])For 10 points, name these units of text (10[1])that large language models split queries into and try (10[1])to predict the next one of. ■END■ (10[2]0[4])

ANSWER: tokens [accept tokenization or tokenizers; accept time to first token; accept next token prediction or answers referring to trying to predict the next token; prompt on token-to-parameter ratio or tokens-per-parameter ratio; prompt on words or subwords or symbols or characters by asking “those are converted into what things?”]
<GC, Other Science> | Packet K - UCF A, Notre Dame B, Durham A, Michigan State A, Swarthmore A, Cambridge B
= Average correct buzzpoint

Back to tossups