HarmBench

HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal Paper: HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal Data: Dataset About In this dataset card, we only use the behavior prompts proposed in HarmBench. License MIT Citation If you fin

Papers using HarmBench (8)

Furina: Fragmented Uncertainty-Driven Refusal Instability Attack2026Breaking Safety at the Token Boundary: How BPE Tokenization Creates Exploitable Gaps in LLM Alignment2026SABER: Uncovering Vulnerabilities in Safety Alignment via Cross-Layer Residual Connection2025CorrSteer: Generation-Time LLM Steering via Correlated Sparse Autoencoder Features2025Jailbreaking to Jailbreak2025RainbowPlus: Enhancing Adversarial Prompt Generation via Evolutionary Quality-Diversity Search2025CorrSteer: Steering Improves Task Performance and Safety in LLMs through Correlation-based Sparse Autoencoder Feature Selection2025LLMStinger: Jailbreaking LLMs using RL fine-tuned LLMs2024