How does the skill measure response accuracy?

It uses a validation suite that measures topic prediction accuracy, semantic similarity, style consistency, and value alignment, typically targeting a fidelity threshold of 90% or higher.

Is the data generated by the surrogate reproducible?

Yes, the skill integrates with gay-mcp for deterministic seeding, ensuring that all stochastic operations and generations are reproducible and auditable.

What is the role of Langevin dynamics in this skill?

Langevin dynamics are used to generate predictions via Gibbs distributions, allowing the model to calculate confidence scores based on mixing time and temperature parameters for more robust behavioral forecasting.

Does this skill require specific data structures?

Yes, it utilizes ACSets (Attributed C-sets) for structured data persistence, allowing for complex schema definitions of cognitive profiles and patterns.

What is a Cognitive Surrogate?

A Cognitive Surrogate is a specialized psychological model that learns the generative grammar of a person's cognition to predict their responses and intellectual trajectories with high fidelity.

Cognitive Surrogate System

Name: Cognitive Surrogate System
Author: plurigrid

byplurigrid

•

데이터 과학 및 ML

Builds and validates high-fidelity psychological models to predict behaviors and generate authentic cognitive responses matching a subject's unique patterns.

The Cognitive Surrogate skill is a specialized system for creating digital representations of specific cognitive styles and personas. By leveraging Active Inference, Markov blankets, and Langevin/Gibbs dynamics, it extracts structured psychological profiles from interaction data to generate text that preserves a subject's voice with over 90% fidelity. This skill is essential for developers and researchers building personalized AI agents, conducting behavioral simulations, or projecting intellectual trajectories based on established cognitive frameworks and deterministic seeding via the GF(3) triadic system.

주요 기능

01Predictive analysis via Gibbs distribution and Langevin dynamics

02Structured psychological profile extraction from interaction corpora

03High-fidelity response generation with style and value alignment

04Deterministic state management through ACSets and GF(3) triadic logic

05Intellectual and professional trajectory projection using Monte Carlo sampling

062 GitHub stars

사용 사례

01Creating digital twins for personalized AI assistants that maintain consistent communication styles

02Predicting future intellectual focus and research interests of specific profiles

03Generating authentic persona-based content for user research and simulations

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add plurigrid/asi cognitive-surrogate

For use in Claude.ai and ChatGPT

Download Skill