Meta-Cognitive Priors
Compact, semantically dense instructions that specify global reasoning properties rather than task content — configuring how a model organizes, weights, and monitors its reasoning before and during task execution. The building blocks of Cognitive Seeds.
Definition
A meta-cognitive prior is a compact, semantically dense instruction that specifies global reasoning properties rather than task content. Instead of telling a model what to think about, a meta-cognitive prior configures how the model organizes, weights, and monitors its reasoning before and during task execution.
Meta-cognitive priors are the building blocks of Cognitive Seeds — the proprietary framework developed by Beau Diamond at NovaThink for reasoning regime activation in frontier language models.
The Problem It Addresses
Standard AI prompts specify content: "Analyze this strategy," "Write copy for this product," "Compare these options." This tells the model what to produce but not how to reason about producing it. The model defaults to its most probable response policy — which is competent but rarely optimal for complex tasks.
Meta-cognitive priors address the process layer. They specify properties like: how many analytical dimensions to sustain, how to weight attention across competing constraints, when to challenge assumptions versus build on them, and how to maintain coherence between different levels of abstraction.
How They Work
Meta-cognitive priors function as process-shaping constraints that bias the model's inference trajectory. By specifying global reasoning properties in semantically dense language, they act as compressed keys that select specific reasoning regimes from the model's learned capabilities.
Their extreme brevity — typically under 30 words — is functionally important. Longer procedural instructions consume the model's finite attention budget with step-by-step procedures, leaving less capacity for the actual reasoning task. A semantically dense prior specifies a global reasoning mode and leaves 99% of the attention budget available for the problem itself. This is why less instruction often produces better reasoning — a concept measured by semantic density.
FAQ
How is a meta-cognitive prior different from a system prompt?
A system prompt typically assigns a role, sets behavioral rules, and specifies output format — all content-level instructions. A meta-cognitive prior specifies reasoning properties: how to allocate attention, how to handle contradictions, how to balance competing constraints. It operates at the process level rather than the content level.
Can anyone create meta-cognitive priors?
The principles behind meta-cognitive priors are accessible to anyone willing to think about AI interaction at the reasoning-process level rather than the content level. Beau Diamond publishes working meta-cognitive priors in his articles and in the Origin Node Zero newsletter for readers to test and learn from.