Large language models can learn and generalize steganographic chain-of-thought under process supervision

Documenti analoghi

• Chain-of-thought obfuscation learned from output supervision can generalise to unseen tasks
  di: Hadida, Nathaniel Mitrani, et al.
  Pubblicazione: (2026)
• Diagnosing Pathological Chain-of-Thought in Reasoning Models
  di: Liu, Manqing, et al.
  Pubblicazione: (2026)
• A transformer architecture alteration to incentivise externalised reasoning
  di: Pavlova, Elizabeth, et al.
  Pubblicazione: (2026)
• Alignment Pretraining: AI Discourse Causes Self-Fulfilling (Mis)alignment
  di: Tice, Cameron, et al.
  Pubblicazione: (2026)
• Setting up for failure: automatic discovery of the neural mechanisms of cognitive errors
  di: Radmard, Puria, et al.
  Pubblicazione: (2025)