Product
Help
Pricing
About
Jobs
Log in
Get started
Keyword
Add to collection
HarmBench
Description:
A benchmark for evaluating language models’ responses to harmful requests and adversarial attacks in AI safety research.
Primary topic:
Adversarial Robustness in Machine Learning (84%)
Related topics:
Adversarial Robustness in Machine Learning (84%),
Ethics and Social Impacts of AI (8%)
OpenAlex ID:
keywords/harmbench
HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal
2024
·
Mazeika, Mantas
, Phan, Long
, et al.
·
arXiv (Cornell University)
·
2,056
PDF
AdvPrompter: Fast Adaptive Adversarial Prompting for LLMs
2024
·
Paulus, Anselm
, Zharmagambetov, Arman
, et al.
·
arXiv (Cornell University)
·
251
PDF
LLM Defenses Are Not Robust to Multi-Turn Human Jailbreaks Yet
2024
·
Li, Nathaniel
, Han, Ziwen
, et al.
·
arXiv (Cornell University)
·
214
PDF
AutoRedTeamer: Autonomous Red Teaming with Lifelong Attack Integration
2025
·
Andy Zhou
, Kevin Wu
, et al.
·
neural information processing systems
·
76
Efficient LLM Jailbreak via Adaptive Dense-to-sparse Constrained Optimization
2024
·
Kai Hu
, Weichen Yu
, et al.
·
neural information processing systems
·
53
Chain-of-Thought Hijacking
2025
·
Zhao, Jianli
, Fu, Tingchen
, et al.
·
arXiv (Cornell University)
·
34
PDF
Advancing Adversarial Suffix Transfer Learning on Aligned Large Language Models
2024
·
Hongfu Liu
, Yuxi Xie
, et al.
·
Conference on Empirical Methods in Natural Language Processing (EMNLP)
·
24
PDF
Will AI Tell Lies to Save Sick Children? Litmus-Testing AI Values Prioritization with AIRiskDilemmas
2026
·
Chiu, Yu Ying
, Wang, Zhilin
, et al.
·
International Conference on Learning Representations
·
22
PDF
AgenticRed: Evolving Agentic Systems for Red-Teaming
2026
·
Yuan, Jiayi
, Nöther, Jonathan
, et al.
·
arXiv (Cornell University)
·
17
PDF
LLMStinger: Jailbreaking LLMs using RL fine-tuned LLMs
2024
·
Jha, Piyush
, Arora, Arnav
, et al.
·
arXiv (Cornell University)
·
17
PDF
Deciphering the Chaos: Enhancing Jailbreak Attacks via Adversarial Prompt Translation
2024
·
Li, Qizhang
, Yang, Xiaochen
, et al.
·
arXiv (Cornell University)
·
17
PDF
Surgery: Mitigating Harmful Fine-Tuning for Large Language Models via Attention Sink
2026
·
Liu, Guozhi
, Lin, Weiwei
, et al.
·
arXiv (Cornell University)
·
9
PDF
LLM Stinger: Jailbreaking LLMs Using RL Fine-Tuned LLMs (Student Abstract)
2025
·
Piyush Jha
, Arnav Arora
, et al.
·
Proceedings of the AAAI Conference on Artificial Intelligence
·
9
PDF
How Sensitive Are Safety Benchmarks to Judge Configuration Choices?
2026
·
Xinran Zhang
·
Communications in computer and information science
·
8
PDF
Do Methods to Jailbreak and Defend LLMs Generalize Across Languages?
2025
·
Atil, Berk
, Passonneau, Rebecca J.
, et al.
·
arXiv (Cornell University)
·
7
PDF
Plentiful Jailbreaks with String Compositions
2024
·
Huang, Brian R. Y.
·
arXiv (Cornell University)
·
7
PDF
How Reliable Is Your Jailbreak Judge? Calibration and Adversarial Robustness of Automated ASR Scoring
2026
·
Gao, Yang
·
arXiv (Cornell University)
·
6
PDF
Jailbreaking in the Haystack
2025
·
Shah, Rishi Rajesh
, Wu, Chen Henry
, et al.
·
arXiv (Cornell University)
·
6
PDF
From Concepts to Components: Concept-Agnostic Attention Module Discovery in Transformers
2026
·
Su, Jingtong
, Kempe, Julia
, et al.
·
International Conference on Learning Representations
·
5
PDF
Cognitive pitfalls of LLMs: a system for generating adversarial samples based on cognitive biases
2024
·
Dong Zhang
, Zhiyuan Hu
, et al.
·
International Conference on Optics, Electronics, and Communication Engineering (OECE)
·
5
Single-Configuration Attack Success Rate Is Not Enough: Jailbreak Evaluations Should Report Distributional Attack Success
2026
·
Maple, Carsten
, Kumar, Abhishek
, et al.
·
arXiv (Cornell University)
·
4
PDF
Dynamic Adversarial Fine-Tuning Reorganizes Refusal Geometry
2026
·
Lan, Wenhao
, Li, Shan
, et al.
·
arXiv (Cornell University)
·
4
PDF
CivicShield: A Cross-Domain Defense-in-Depth Framework for Securing Government-Facing AI Chatbots Against Multi-Turn Adversarial Attacks
2026
·
Patil, KrishnaSaiReddy
·
arXiv (Cornell University)
·
4
PDF
FreakOut-LLM: The Effect of Emotional Stimuli on Safety Alignment
2026
·
Daniel Kuznetsov
, Ofir Cohen
, et al.
·
4
PDF
Membrane: A Self-Evolving Contrastive Safety Memory for LLM Agent Defense
2026
·
Minseok Choi
, Seungbin Yang
, et al.
·
Conference on Empirical Methods in Natural Language Processing (EMNLP)
·
4
PDF
Show more works
View as search
Works count:
132
Citation count:
2,901
Year
Topic
Type