Beyond Next Token Probabilities: Learnable, Fast Detection of Hallucinations and Data Contamination on LLM Output Distributions

Documenti analoghi

• Beyond Token Probes: Hallucination Detection via Activation Tensors with ACT-ViT
  di: Bar-Shalom, Guy, et al.
  Pubblicazione: (2025)
• Neural Message-Passing on Attention Graphs for Hallucination Detection
  di: Frasca, Fabrizio, et al.
  Pubblicazione: (2025)
• On The Expressive Power of GNN Derivatives
  di: Eitan, Yam, et al.
  Pubblicazione: (2025)
• Topological Blindspots: Understanding and Extending Topological Deep Learning Through the Lens of Expressivity
  di: Eitan, Yam, et al.
  Pubblicazione: (2024)
• A Flexible, Equivariant Framework for Subgraph GNNs via Graph Products and Graph Coarsening
  di: Bar-Shalom, Guy, et al.
  Pubblicazione: (2024)