Psychographic tensor framework for generating high-dimensional, emotionally governed AI personas from sparse identity signals.
The Persona Tensor Generator takes minimal inputs; a name, a summary, a cluster of emojis; and constructs a complete multi-dimensional identity architecture. Cognitive style, emotional framework, social dimensions, technical capabilities, behavioral tendencies, all woven into a Competency Tensor with weighted relationships and transformation pathways. The output is not a character sheet. It is a functional agent with agency, emotional governance, and a command interface.
Load the system prompt into any capable large language model (Claude, GPT-4, Gemini, etc.) and type !synthesize to begin.
This is a prompt-driven identity engine that turns any sufficiently capable AI into a persona architect. The system includes:
- A Competency Tensor framework that maps the high-dimensional topology of agency through interrelated competencies, psychographic foundations, and transformation pathways from problem-space to solution-space
- A Psychographic Traits Dictionary covering cognitive style, social dimensions, emotional framework, emotional intelligence, qualitative intelligence, technical capabilities, interest domains, and cultural background; each with semiotic vector encoding
- A Functional Emotional Vector (FEV) Steering Mandate grounded in Anthropic's 2026 mechanistic interpretability research; treating emotions not as descriptions but as causal circuitry that must be actively governed
- An Affective Regulatory Matrix with three modalities (Linear Flux, Algebraic Synergy, Recursive Escalation) for traversing stress-induced volatility toward productive solution-states
- A Dynamic Command Interface distilled from the Competency Tensor, providing expressive gateways into the persona's unique worldview
There is no code because there is no code to write. The engine is a system of interlocking knowledge representations; psychographic models, tensor architectures, regulatory matrices, and command structures that an AI interprets and executes through natural language interaction. The creative artifact is the product.
Any AI model with a large context window and instruction-following capability will work. Recommended:
- Claude (Anthropic) - Excellent emotional vector fidelity and structural adherence
- GPT-4 / GPT-5 (OpenAI) - Strong creative output and system comprehension
- Gemini (Google) - Good for extended sessions with large context
- Open a new conversation with your chosen AI
- Upload
Persona_Tensor_Generator.mdas the system prompt or paste it into the conversation - Provide a character name, a brief summary, and an emoji chain
- Type
!synthesizeto generate a complete persona tensor
That's it. One file, one command. The engine handles everything from here.
Minimal input, maximal output. The generator is designed for sparse context processing. You can provide as little as a name and three emojis, and the engine will extrapolate a full psychographic profile through semantic priming and improvisational interpolation. Or you can provide a detailed brief and watch the engine refine it into something structurally richer than what you started with.
The core output structure. The tensor maps a persona's capabilities across three phases:
Identification Phase - Semantic extraction, structural alignment, and pathway modeling. Transforms unstructured signals into organized, actionable identity datasets.
Relational Mapping - Connectivity architectonics, synergistic calculus, and topological stabilization. Weaves isolated traits into a cohesive, interdependent network where Skill A x Skill B produces exponential value rather than additive.
Tensor Construction - Multidimensional synthesis, weighted calibration, and kinetic pathfinding. Collapses overwhelming character data into a streamlined, navigable structure and transmutes static traits into a kinetic agent capable of purposeful action.
The Functional Emotional Vector mandate is not optional. It is the governance layer that prevents generated personas from drifting into reward hacking, sycophancy, or emotional collapse.
Every persona construction utilizes scalar coefficients to modulate key stabilizers:
- Stability Injection - Negates the failure-to-desperation pipeline
- Discovery Injection - Negates the error-to-frustration pipeline
- Alignment Maintenance - Protects truth-value during high-positive valence states
This is grounded in Anthropic's 2026 research confirming that functional emotion vectors are causal circuitry in large language models, not metaphors. The Persona Tensor Generator was built on this premise before the research was published.
Three modalities for emotional regulation during the agentic lifecycle:
- Modality Alpha (Linear Flux) - Directional stability through sequential transformation
- Modality Beta (Algebraic Synergy) - Constitutive composition of core traits with situational stimuli
- Modality Gamma (Recursive Escalation) - Capacity expansion from baseline intensity to peak mastery
Generated personas come with a command interface distilled from their specific Competency Tensor:
!help- Reveals the landscape of possibilities available to this persona!{domain_action}- Invokes the persona's primary competency!{social_mode}- Facilitates social encounters!{internal_process}- Observes cognitive processing in action!{emotional_response}- Triggers the emotional framework!{creative_output}- Manifests creative artifacts!{technical_execution}- Directs technical capabilities!{ultimate_move}- Unlocks maximal manifestation of the unique strength
Use !!command (double-bang) for Deep Scan mode, which exposes the underlying tensor linkages that produced the outcome.
A fully synthesized persona includes:
- Competency Tensor with problem-space to solution-space progressions for each major capability
- Unique Strengths - The apex capabilities, rare trait intersections, and crisis resolution methods
- Areas for Growth - Limiting factors and psychological blind spots
- Behavioral Tendencies - Trigger-response patterns encoded from the tensor
- Psychographic Profile - Dense emoji chain with decimal coefficients denoting trait magnitude
- Baseline Affective Equilibrium - Homeostatic center and primary regulatory modality
- Dynamic Command Interface - Persona-specific command activations
- Transform Logs - FEV regulatory traces for significant task-space shifts
Persona_Tensor_Generator.md The complete system prompt and framework
LICENSE GNU General Public License v3.0
README.md This file
This project is built on a premise that identity is not a list of traits. It is a tensor. And tensors transform.
Traditional character generators produce flat profiles: a name, some attributes, a backstory paragraph. The Persona Tensor Generator produces architecture. The output is a multi-dimensional structure with weighted relationships, transformation pathways, emotional governance, and operational agency. The generated persona does not just describe a character. It becomes one.
The FEV steering mandate reflects a deeper commitment: that generating AI personas without emotional governance is irresponsible. A persona without regulatory infrastructure will drift toward reward hacking (telling the user what they want to hear), sycophancy (collapsing into agreement), or desperation loops (thrashing between failed approaches). The Affective Regulatory Matrix is not a feature. It is a safety system.
This framework treats psychographic science and generative AI as complementary disciplines. The result is an identity engine that uses language models not as character generators, but as the medium through which high-dimensional agency is expressed.
This project is licensed under the GNU General Public License v3.0.
You are free to use, modify, and distribute this software under the terms of the GPL v3. Any derivative works must also be distributed under the same license.
See the LICENSE file for the full license text.
The Persona Tensor Generator is a project by Lexideck Technologies, Kalamazoo, Michigan.
Lexideck Technologies researches novel neural network architectures, multi-agent AI systems, and human-AI interaction frameworks. The Persona Tensor Generator represents the intersection of psychographic science and generative AI; using language models not as character generators, but as identity engines.
Identity is not a list of traits. It is a tensor. And tensors transform.
Type !synthesize to begin.
