Set a Specific IQ Rating to Control Writing Analysis Depth

Assigning a numeric IQ primes the model with a concrete persona anchor. Rather than defaulting to a generic 'expert' frame, the model maps the number onto associated traits — analytical rigor, abstraction level, vocabulary — and calibrates its output accordingly. It's a shorthand for a rich bundle of behavioral expectations. The scale effect the author describes is real: higher numbers push the model toward more abstract reasoning, obscure rhetorical frameworks, and deeper structural critique. Lower numbers produce competent but conventional feedback. The number acts like a resolution dial on how far the model reaches into its training for rare or advanced concepts. This works because LLMs respond to role-setting prompts that give them a measurable, internally consistent reference point. A vague instruction like 'be smart' is under-constrained; a number like 145 or 160 is specific enough to trigger a coherent persona shift across the response.