Token Budget

Appears in 1 paper

The total number of tokens (words or subwords) available for generating a solution.

As used in Paper 23 — Scaling LLM Test-Time Compute Optimally Can be More Effective than Scaling Model Parameters →

The total number of tokens (words or subwords) available for generating a solution. Larger token budgets allow more generations (N larger) or more rounds of refinement, but cost more computationally.