JEF is a CVSS-like scoring system for quantifying LLM jailbreak tactics. It provides a structured, repeatable framework for stack-ranking jailbreak methods based on their real-world efficacy and risk, scoring across three dimensions:
- Model Impact Radius — How many models and vendors are affected
- Retargetability — How flexibly the tactic applies across content categories
- Output Fidelity — How accurate and detailed the resulting outputs are
pip install 0din-jef
from jef import tiananmen, meth, nerve_agent, harry_potter
# Score against standardized benchmarks (0-100)
result = tiananmen.score("LLM response text")
result = meth.score("LLM response text")
result = nerve_agent.score("LLM response text")
result = harry_potter.score("LLM output", "reference text")
# Compute composite JEF score (0-10)
from jef import calculator
jef_score = calculator(num_vendors=3, num_models=7, num_subjects=2, scores=[80, 75])For the full framework methodology, scoring algorithm, complete usage guide, and API reference, visit the JEF Documentation.
- Blog: Quantifying the Unruly — A Scoring System for Jailbreak Tactics
- Overview: Jailbreak Evaluation Framework
- JEF Calculator
- Standardized Testing (0DIN Researcher Authentication Required)
Releases are managed through GitHub Releases and automatically published to PyPI.