Welcome to promptbench’s documentation!

PromptBench is a unified library for evaluating and understanding large language models. Please refer to PromptBench for the code.

Get Started

• promptbench Introduction
• Installation

Examples

• Basic Usage
• Multi-Modal Models
• DyVal Evaluation
• Prompt Attack
• Prompt Engineering
• Adding new modules

Reference

• dataload
• dyval
• metrics
• models
• utils

Leaderboards

• Adversarial Prompt Leaderboard
• Dynamic Evaluation Benchmark
• Prompt Engineering Benchmark

Indices and tables

• Index

• Module Index

• Search Page