When "Competency" in Reasoning Opens the Door to Vulnerability: Jailbreaking LLMs via Novel Complex Ciphers

Documenti analoghi

• GuidedSampling: Steering LLMs Towards Diverse Candidate Solutions at Inference-Time
  di: Handa, Divij, et al.
  Pubblicazione: (2025)
• ThinkTuning: Instilling Cognitive Reflections without Distillation
  di: RRV, Aswin, et al.
  Pubblicazione: (2025)
• ActionReasoningBench: Reasoning about Actions with and without Ramification Constraints
  di: Handa, Divij, et al.
  Pubblicazione: (2024)
• Chaos with Keywords: Exposing Large Language Models Sycophantic Hallucination to Misleading Keywords and Evaluating Defense Strategies
  di: RRV, Aswin, et al.
  Pubblicazione: (2024)
• Hypothesis Generation for Materials Discovery and Design Using Goal-Driven and Constraint-Guided LLM Agents
  di: Kumbhar, Shrinidhi, et al.
  Pubblicazione: (2025)