Skip to main content

Interactive and Iterative Prompt Mutation Interface for LLM Developers and Evaluators

Project description

Muteract PyPI version License

Interactive and Iterative Prompt Mutation Interface for LLM Developers and Evaluators Screenshot of the Muteract User Interface

Table of Contents

Introduction

Muteract - an interactive and iterative prompt mutation interface that enables LLM developers and evaluators to input natural language (NL) text prompts, apply mutations, analyze variations in textual responses, and archive results.

structure

As of now, this tool provides only Radamsa as the mutator, since it works directly on the bytes in a prompt and can be applied to various modalities. We plan to add more mutators for specific to images, text etc. in the future.

The interaction flow of Muteract is

flowchart

Getting Started

Muteract is a python based tool. Make sure python is installed before following the Installation guide.

Installation

Muteract can be installed with a simple pip command.

# Installation command
pip install muteract

All the dependencies are taken care by the installation.

Usage

Ensure that the OpenAI API Key is configured in the environment variable OPEN_AI_API_KEY before starting the application.

Just running the Muteract command will open the GUI:

muteract

Configuration

Python version above 3.10 is needed for running the application, along with a browser that supports ES2017.

Acknowledgements

This tool is being developed by SET-IITGN Group in collaboration with HAIx Lab, IITGN.

This work is supported by Grant No. IP/IITGN/CSE/SM/2324/02 and Grant No. IP/IITGN/CSE/YM/2324/05 from IIT Gandhinagar, Gujarat, India.

Contributing

Conrtibutions are accepted via pull requests. The PRs will be accepted only if they are suitable for the tool.

License

Apache License

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

muteract-0.2.0.tar.gz (692.5 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

muteract-0.2.0-py3-none-any.whl (709.1 kB view details)

Uploaded Python 3

File details

Details for the file muteract-0.2.0.tar.gz.

File metadata

  • Download URL: muteract-0.2.0.tar.gz
  • Upload date:
  • Size: 692.5 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.1.0 CPython/3.12.1

File hashes

Hashes for muteract-0.2.0.tar.gz
Algorithm Hash digest
SHA256 2c563b3242a95a55abcee178e0710f1c532a2560ac1db99d60ed3f50afd8ed21
MD5 941f595473983efad6ad82a42f926b88
BLAKE2b-256 4c3e773d643f5f7088be59e4ea411bf51fb77eab6c82377a42691535c6b1a9aa

See more details on using hashes here.

File details

Details for the file muteract-0.2.0-py3-none-any.whl.

File metadata

  • Download URL: muteract-0.2.0-py3-none-any.whl
  • Upload date:
  • Size: 709.1 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.1.0 CPython/3.12.1

File hashes

Hashes for muteract-0.2.0-py3-none-any.whl
Algorithm Hash digest
SHA256 7b7894de917a28827eb9e33f4254542e447698e2567de1a117a2955d71bf3a7e
MD5 55d8f94cee9cd3ad13472856da249f57
BLAKE2b-256 81cc88b595dfb992277efe36b978d68621540197322bdf60ac1aae035407bb67

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page