Latent Phase-Shift Rollback: Inference-Time Error Correction via Residual Stream Monitoring and KV-Cache Steering

↘ Related papers

• Knowing When to Quit: A Principled Framework for Dynamic Abstention in LLM Reasoning
  Hen Davidov48% match
• Too Correct to Learn: Reinforcement Learning on Saturated Reasoning Data
  Zhenwen Liang43% match
• Semantic Step Prediction: Multi-Step Latent Forecasting in LLM Reasoning Trajectories via Step Sampling
  Yidi Yuan39% match
• When Can LLMs Learn to Reason with Weak Supervision?
  Salman Rahman36% match
• Train Separately, Merge Together: Modular Post-Training with Mixture-of-Experts
  Jacob Morrison35% match