---
title: PHANTASM LLM Hallucination Inverter
emoji: 🔮
colorFrom: purple
colorTo: indigo
sdk: gradio
sdk_version: 5.9.1
app_file: app.py
pinned: true
license: apache-2.0
tags:
  - hallucination
  - llm
  - uncertainty
  - confabulation
  - calibration
  - nlp
  - ai-safety
  - phantasm
  - gradio
short_description: Invert LLM hallucination into productive features
---

# 🔮 PHANTASM Demo

**Probabilistic Hallucination-Aware Neural Transformation with Adaptive Synthesis Method**

The first ML framework to mathematically invert LLM failure modes into productive features.

## What this demo does

Paste any text to get a full three-pillar PHANTASM analysis:

- **Pillar I — HGT**: Hallucination risk score + knowledge boundary tokens
- **Pillar II — CMN**: Novel hypotheses mined from confabulation patterns  
- **Pillar III — UC**: Calibrated confidence + reliability tier (crystal/solid/fluid/vapor)

## Links

- [GitHub](https://github.com/vignesh2027/PHANTASM)
- [PyPI](https://pypi.org/project/phantasm-llm/)
- [Dataset](https://huggingface.co/datasets/vigneshwar234/PHANTASM-benchmark)
- [Docs](https://vignesh2027.github.io/PHANTASM)