Parks staff keep getting overlong brochure drafts from their generative app. Which model-parameter change best addresses that?

Parks staff getting overlong brochure drafts should adjust max tokens and stop sequences to constrain draft length. Max tokens caps how much the model can emit; stop sequences end generation at structural boundaries (section end, flyer length, delimiter). Together they enforce brochure-scale output instead of epic sprawl. That is the direct model-parameter fix for length, not a change to temperature, top-p, or presence penalty. Increasing temperature to produce more focused concise brochure draft wording fails because temperature controls randomness and creativity, not length or focus; it won't enforce brevity and can make output less predictable. Adjusting top-p sampling to directly limit brochure draft length fails because top-p controls the diversity of token choices, not a hard cap on output length; it cannot stop a draft from running long. Raising presence penalty to enforce brochure section endings cleanly fails because presence penalty discourages repeated tokens; it does not create section endings or constrain overall draft length. Exam caveat: length control for generative brochure drafts is primarily max tokens and stop sequences—not raising temperature, adjusting top-p, or raising presence penalty. Operational check: set brochure-appropriate max tokens, define stop sequences at flyer boundaries, regenerate sample parks drafts, and confirm average length stays within the published flyer template before releasing the parameter change.