transformer-bottleneck

#transformer-bottleneck

When Chain-of-Thought Helps and When It Hurts: An Empirical Investigation of the Serial-Depth Bottleneck in LLM Reasoning

arXiv cs.CL ↗ · 2026-08-12 Cached

This empirical study tests when chain-of-thought prompting helps or hurts LLM reasoning, finding that CoT provides large gains on deep serial tasks like GSM8K and MATH but is redundant on shallow tasks like MMLU and ARC-Challenge, consistent with a serial-depth bottleneck framework.

0 favorites 0 likes

transformer-bottleneck

When Chain-of-Thought Helps and When It Hurts: An Empirical Investigation of the Serial-Depth Bottleneck in LLM Reasoning

Submit Feedback