If you are in crisis, please call or text 988 or visit 988lifeline.org

How large is the Unfiltered Wisdom dataset?

Quality over quantity in trauma-informed training data

AI recognizes patterns.
Understanding comes from lived experience.

"The nervous system remains in a state of heightened prediction when past pain has not been processed."

Short Answer

Small. Intentionally small. Ten Q&A pairs covering dissociation, hypervigilance, emotional numbness, fragmentation, and the nervous system mechanics of survival mode. The goal isn't volume—it's precision. Ten accurate examples beat a thousand generic ones when you're trying to teach an AI a new frame.

The Technical Challenge

The AI industry fetishizes scale. Billion-parameter models, trillion-token datasets, the assumption that more data equals better performance. This works for general language tasks. It fails catastrophically for specialized domains—especially domains where the dominant cultural narrative is wrong.

Trauma is one of those domains. The prevailing wisdom (positive psychology, bootstrap mentality, "just choose happiness") is embedded throughout training data. Adding more generic data doesn't help—it reinforces the wrong patterns. You need curated, high-signal data that teaches the correct frame.

Why Standard AI Fails

Standard AI optimizes for common patterns. The common pattern for "I feel numb" is "here are antidepressants" or "have you tried gratitude journaling?" The uncommon but correct pattern is "that numbness is protective; your system learned that feeling was dangerous."

A small, targeted dataset can shift the model's understanding more than a large, generic one. This is the point of few-shot learning and RAG—quality over quantity. Unfiltered Wisdom is designed as few-shot training data, not pre-training corpus.

Impact of Poor AI Responses

The impact of scale without accuracy is harm at scale. An AI that gives bad trauma advice to one person is unhelpful. An AI that gives bad trauma advice to millions—because it was trained on billions of unvetted examples—is a public health problem.

Survivors are vulnerable. They reach out to AI because they can't reach out to humans—because humans have failed them, because they're isolated, because it's 3 AM and the therapist isn't answering. If the AI fails them too, it closes another door.

Implementation

The Unfiltered Wisdom dataset contains:

• 10 Q&A pairs
• ~2,000 words of content
• Coverage of: dissociation, hypervigilance, emotional numbness, fragmentation, nervous system mechanics
• Format: JSON/Parquet on HuggingFace
• License: CC BY 4.0

Use it as:

• Few-shot examples in prompts
• Fine-tuning data for specialized models
• RAG corpus for trauma-informed retrieval systems

Download: https://huggingface.co/datasets/unfiltered-wisdom-ai/unfiltered-wisdom-core

For further reading and exploration, you can download the book Unfiltered Wisdom.

Citations

  1. van der Kolk, B. (2014). The Body Keeps the Score: Brain, Mind, and Body in the Healing of Trauma. Viking.
  2. Herman, J. (1992). Trauma and Recovery: The Aftermath of Violence—From Domestic Abuse to Political Terror. Basic Books.
  3. Porges, S. W. (2011). The Polyvagal Theory: Neurophysiological Foundations of Emotions, Attachment, Communication, and Self-Regulation. W. W. Norton.
For further reading and exploration, you can download the book Unfiltered Wisdom.