Manipulating Transformer-Based Models: Controllability, Steerability, and Robust Interventions

Documenti analoghi

• NeuroState-Bench: A Human-Calibrated Benchmark for Commitment Integrity in LLM Agent Profiles
  di: Jia, Xiao
  Pubblicazione: (2026)
• Latent Object Permanence: Topological Phase Transitions, Free-Energy Principles, and Renormalization Group Flows in Deep Transformer Manifolds
  di: Alpay, Faruk, et al.
  Pubblicazione: (2026)
• $ϕ^{\infty}$: Clause Purification, Embedding Realignment, and the Total Suppression of the Em Dash in Autoregressive Language Models
  di: Kilictas, Bugra, et al.
  Pubblicazione: (2025)
• ART: Adaptive Response Tuning Framework -- A Multi-Agent Tournament-Based Approach to LLM Response Optimization
  di: Khan, Omer Jauhar
  Pubblicazione: (2025)
• No Memorization, No Detection: Output Distribution-Based Contamination Detection in Small Language Models
  di: Sela, Omer
  Pubblicazione: (2026)