hypogenic - Generate and test hypotheses from tabular data
Generates and tests hypotheses from observational tabular datasets, optionally integrating research literature through HypoGeniC, HypoRefine, and Union methods.
Tags
Updated: 2026-10-16Capabilities
Typical Inputs
Typical Outputs
What this skill does
- Generate data-driven hypotheses
- Refine hypotheses iteratively
- Test competing hypotheses
- Integrate research literature
- Process research PDFs
- Run hypothesis inference
- Cache LLM responses
- Process datasets in parallel
Inputs
- Tabular datasets
- Task configuration YAML
- Prompt templates
- Research paper PDFs
- Hypothesis bank files
- Test dataset paths
Outputs
- Generated hypothesis files
- Hypothesis inference results
- Processed literature files
- Validation performance
Requirements
- Hypogenic package
- HuggingFace-format datasets
- API-based or local LLM
- Configured task prompts
- GROBID for PDF preprocessing
- s2orc-doc2json for literature processing
- Redis server for caching