Signal2026-07-16
Papers With Code

Length Penalties Make Chain-of-Thought Less Monitorable

Part of

LLM Self-Distillation For Enhanced Reasoning

Open primary sourceSee the whole picture