Cheating Automatic LLM Benchmarks: Null Models Achieve High Win Rates

Documenti analoghi

• Improved Few-Shot Jailbreaking Can Circumvent Aligned Language Models and Their Defenses
  di: Zheng, Xiaosen, et al.
  Pubblicazione: (2024)
• Agent Smith: A Single Image Can Jailbreak One Million Multimodal LLM Agents Exponentially Fast
  di: Gu, Xiangming, et al.
  Pubblicazione: (2024)
• Improving Your Model Ranking on Chatbot Arena by Vote Rigging
  di: Min, Rui, et al.
  Pubblicazione: (2025)
• Denial-of-Service Poisoning Attacks against Large Language Models
  di: Gao, Kuofeng, et al.
  Pubblicazione: (2024)
• Benchmarking Large Multimodal Models against Common Corruptions
  di: Zhang, Jiawei, et al.
  Pubblicazione: (2024)