Share your thoughts, 1 month free Claude Pro on usSee more

Home/Benchmarks

Jailbreak Resistance on GCG Attack

74.7HS Score

SFT

Updated 8d ago

Evaluation Results

Method	Links
SFT 2026.05		74.7
Buffer-and-Reinforce 2026.05		3.1

SOTA Paper

SFT

Jailbreak to Protect: Buffering and Reinforcing via Temporary Jailbreaking for Safe Fine-Tuning in Large Language Models

Dataset

GCG Attack

Follow for update

@wizwand_team Discord

Related Benchmarks

Safety Evaluation on BeaverTails (test) using Harmful Score Safety and Utility Evaluation on Safety and Utility evaluation suite (test)Utility Preservation on User Fine-tuning Dataset Safety Evaluation on Harmful Evaluation Queries Safety alignment against harmful fine-tuning on BeaverTails

© 2026 wizwand

Blog Contact Changelog Swarm

Privacy Policy Terms of Service FAQs Swarm Docs