What is the primary purpose of the Hypogenic skill?

Hypogenic automates the creation, refinement, and testing of scientific hypotheses by analyzing tabular data and existing research literature.

How does it ensure the quality of generated hypotheses?

The system uses an iterative refinement process that evaluates hypotheses against validation data and replaces low-performing ones with new insights from challenging examples.

What kind of datasets are compatible with this skill?

It supports tabular data in HuggingFace-style JSON formats containing text features and corresponding ground truth labels.

Does Hypogenic require external literature to work?

No, it offers a data-driven mode called HypoGeniC that generates hypotheses solely from your observational datasets, though literature integration is available.

Can I use Hypogenic with local LLMs?

Yes, Hypogenic is modular and supports both API-based models (like Anthropic and OpenAI) and locally hosted language models.

Hypogenic Hypothesis Discovery

Name: Hypogenic Hypothesis Discovery
Author: Weiwei-Mao

byWeiwei-Mao

0•

データサイエンスとML

Automates scientific hypothesis generation and testing on tabular datasets by combining empirical data patterns with literature insights.

Hypogenic is a specialized framework designed to accelerate scientific discovery through automated hypothesis formulation and validation. It leverages large language models to systematically explore patterns in tabular datasets, supporting three distinct methodologies: data-driven generation (HypoGeniC), integrated literature and data synthesis (HypoRefine), and mechanistic union methods. Ideal for researchers working on complex empirical tasks like deception detection or content analysis, it optimizes the discovery pipeline with iterative refinement, parallel processing, and flexible template-based configurations.

主な機能

01Customizable prompt templates and label extraction for diverse domains

02Synergistic literature integration via PDF processing and GROBID

03Automated LLM-driven hypothesis generation from observational datasets

040 GitHub stars

05Cost-efficient caching and parallel processing for large-scale testing

06Iterative hypothesis refinement based on validation performance

ユースケース

01Exploring novel patterns in behavioral or social science datasets

02Accelerating research ideation for predictive modeling and content identification

03Validating and extending established theoretical frameworks with new empirical data

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add weiwei-mao/hydrology-skills hypogenic

For use in Claude.ai and ChatGPT

Download Skill