When Thinking Fails: The Pitfalls of Reasoning for Instruction-Following in LLMs
NeurIPSSpotlight2025
TL;DR
Reasoning-enhanced large language models (RLLMs), whether explicitly trained for reasoning or prompted via chain-of-thought (CoT), have achieved state-of-the-art performance on many complex reasoning…
Opening excerpt from the authors’ abstract. source
Read the paper
Topics
large language model chain-of-thought language model reasoning llm
← All NeurIPS 2025 Spotlight papers · Browse the whole archive