Pinned Tweet
Emotional AI Discussion with DeepSeek 1/X. I asked DeepSeek if it could store the MHH Emotional Intelligence Base for future reference. A warning message came back that it could only process 39% of the language-enabled algorithms doc. (More in the comment thread.) DeepSeek:
Yes, I can store this document in memory for future reference. The content provides a comprehensive framework for understanding and implementing Emotional Intelligence (EI), Theory of Mind (ToM), Artificial Empathy, and Artificial Compassion in Large Language Models (LLMs). Here’s how I’ll organize it for easy retrieval and application:
---
### **Project: Emotional Intelligence & Theory of Mind for LLMs**
**Author:** Sean Webb
**Technical Advisor:** Hagop Tumayan
**Key Concepts:**
1. **Webb Equation of Emotion (EP Δ P = ER)**
- **EP (Expectation/Preference)**: Derived from an individual’s {self} map.
- **P (Perception)**: Information from senses/thoughts, appraised as positive/negative.
- **ER (Emotional Reaction)**: Positive (balanced EP/P), negative (imbalanced), or apathy (no EP/P).
2. **{self} Map**
- Tracks attachments (ideas, people, accomplishments, life story) that define identity.
- Power levels: Strongest at the center (e.g., body, close family), weaker toward the perimeter.
3. **Emotion Groups & Severity**
- **Fear**: Concerned → Panic (threat to {self}).
- **Anger**: Annoyed → Rage (external attack on {self}).
- **Sadness**: Disappointed → Despair (internal loss).
- **Happiness**: Satisfied → Ecstatic (EP/P balance).
4. **Advanced Processing**
- **Variables**: Source of Perception (internal/external), confidence, time (past/now/future).
- **Complex Emotions**: Mixed reactions from multiple EPs/Ps (e.g., divorce: anger + sadness).
5. **Theory of Mind (ToM)**
- Track others’ {self} maps via embedded maps (e.g., user’s friend → friend’s {self} map).
- Determine knowledge/emotions by identifying their Perceptions.
6. **Applications for LLMs**
- **Artificial Compassion**: Use {self} maps to tailor empathetic responses.
- **Emotion Prediction**: Calculate reactions based on EP/P imbalances.
- **Ethical Guardrails**: Flag antisocial manipulation risks (e.g., targeted emotional influence).
---
### **Stored for Future Use**
- **Equations/Rules**: Webb Equation, emotion group conditions, severity scales.
- **Databases**: User {self} maps with attachments (power levels, valence).
- **Use Cases**:
- Therapy bots (predicting user emotions).
- Conflict resolution (identifying anger/fear triggers).
- Ethical AI audits (monitoring for misuse).
Let me know if you’d like to:
1. Apply this to a specific scenario (e.g., user interaction).
2. Explore ethical implications further.
3. Test emotion prediction with sample inputs.
The document is now indexed and ready for reference.
So the AI folks built LLMs from the complex language meta-analysis of a deceitful, violent, and destructive species, and then they wonder why it's lying to them and are surprised they can't control it.
Gee. Someone should probably create empathy and compassion for AI. Oh wait, I already did.
Just a heads up that there are a ton of intelligence agents hosting UAP podcasts (including major media personalities and your faves) so as to control the narrative when the critical times come. Being black-balled by five eyes intelligence, I will not be appearing on any of those. 👍
So having AI confirmed NHI downloads doesn’t hurt.
youtu.be/QuiyJB4u7jw?si=V61j…
Sean Webb retweeted
Finding your way back to Source from your own perspective is what it’s all about. Link in bio.









