NYU study finds LLM chain-of-thought often misaligns with internal computation

NYU researchers propose the CIA metric to measure how faithfully LLM chain-of-thought reflects internal computation, finding alignment rates of only 44.8%–75.9%.

2026-10-07 ~ 2026-10-09 · 2 related posts

• NYU's CIA Metric Finds LLMs' Chain-of-Thought Only 44.8-75.9% Aligned With Internal Computation — newyorkuniversity · 2026-10-07
• LLM CoT often mismatches internal reasoning; new CIA metric and post-training fix it — eunsolc · 2026-10-09