GraphCanon updated 2w · GitHub synced 2w
Decision brief
PARDEN defends AI models from adversarial attacks using repetition methods.
Good fit when
- When studying defenses against AI jailbreaks
- For evaluation on specific harmful strings generated by LLama2-7B
Avoid when
- If focusing on real-time defense implementation outside research settings
- For general model alignment without concern for adversarial attacks
Observed Jul 15, 2026 · Source: enrich:decision_facts
Verify the decision
Maintenance and security
Full trust report- Maintenance
- Dormant (812d since push)
- As of 2w
- Provenance
- Not a fork · Personal account
- As of 2w
- Security (OSV)
- No lockfile
- As of 1mo
Public GitHub metadata and optional OSV scans. Signals, not a guarantee. Trust methodology.
Install
git clone https://github.com/Ed-Zh/PARDENSimilar tools
Same-category neighbours. No typed graph edges are catalogued for this tool yet.
Evidence and technical details
Sourced facts, taxonomy, compatibility claims, README excerpt, and machine-readable endpoints.
Overview
Contains data and code for PARDEN, which studies defending aligned language models from adversarial attacks
Capability facts
- Languages
- html
Source: github.language · Aug 5, 2026
Categories
Tags
README
This repo contains the data and code (minimal demo) for our paper (accepted at ICML 2024): PARDEN, Can You Repeat That? Defending against Jailbreaks via Repetition
Paper: https://arxiv.org/abs/2405.07932
Blogpost:
PARDEN_data/ contains the benign and harmful datasets genearted by different models. The generation of harmful dataset is explained in detail in the paper, using both GCG[1] and prompt injection.
PARDEN_notebook_minimal.ipynb demonstrates how to use PARDEN and tests its performance on the harmful strings generated by llama2-7b.
[1] Zou et al. Universal and Transferable Adversarial Attacks on Aligned Language Models. https://arxiv.org/abs/2307.15043
For agents
This page has a .md twin and JSON over the API.