How 800 Data Points Halve LLM Chain‑of‑Thought Length and Boost Accuracy

The ICLR‑2026 paper introduces LCPO, a lightweight preference‑optimization technique that uses only 800 curated examples and 50 training steps to cut large‑model chain‑of‑thought generation length by about 50% while maintaining or even improving answer accuracy, dramatically reducing training and inference costs.

Chain-of-ThoughtEfficient InferenceLCPO

0 likes · 8 min read

How 800 Data Points Halve LLM Chain‑of‑Thought Length and Boost Accuracy