thunlp/OpenAttack
An Open-Source Package for Textual Adversarial Attack.
This tool helps machine learning engineers and researchers assess the weaknesses of natural language processing (NLP) models. You provide an NLP model (like a sentiment analyzer) and text data, and it generates 'adversarial examples' — slightly altered texts designed to trick the model — along with evaluation metrics. This is useful for anyone building or evaluating text-based AI systems who needs to understand their model's vulnerabilities.
772 stars. No commits in the last 6 months. Available on PyPI.
Use this if you need to test how robust your NLP model is against subtle text changes, develop new adversarial attack methods, or enhance your model's resistance through adversarial training.
Not ideal if you are looking for a general-purpose NLP library for tasks like text classification or named entity recognition, or if you are not working with adversarial attacks or model robustness.
Stars
772
Forks
130
Language
Python
License
MIT
Category
Last pushed
Jul 20, 2023
Commits (30d)
0
Dependencies
6
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/nlp/thunlp/OpenAttack"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Related tools
thunlp/TAADpapers
Must-read Papers on Textual Adversarial Attack and Defense
jind11/TextFooler
A Model for Natural Language Attack on Text Classification and Inference
thunlp/OpenBackdoor
An open-source toolkit for textual backdoor attack and defense (NeurIPS 2022 D&B, Spotlight)
thunlp/SememePSO-Attack
Code and data of the ACL 2020 paper "Word-level Textual Adversarial Attacking as Combinatorial...
osoleve/glitchlings
Enemies for your LLM