EVOREFUSE: Evolutionary Prompt Optimization for Evaluation and Mitigation of LLM Over-Refusal to Pseudo-Malicious Instructions

Documenti analoghi

• TRIDENT: Enhancing Large Language Model Safety with Tri-Dimensional Diversified Red-Teaming Data Synthesis
  di: Wu, Xiaorui, et al.
  Pubblicazione: (2025)
• Are Emotion and Rhetoric Neurons in LLM? Neuron Recognition and Adaptive Masking for Emotion-Rhetoric Prediction Steering
  di: Zheng, Li, et al.
  Pubblicazione: (2026)
• DiscoSG: Towards Discourse-Level Text Scene Graph Parsing through Iterative Graph Refinement
  di: Lin, Shaoqing, et al.
  Pubblicazione: (2025)
• Revisiting Structured Sentiment Analysis as Latent Dependency Graph Parsing
  di: Zhou, Chengjie, et al.
  Pubblicazione: (2024)
• CMNER: A Chinese Multimodal NER Dataset based on Social Media
  di: Ji, Yuanze, et al.
  Pubblicazione: (2024)