What Affects the Effective Depth of Large Language Models?

TL;DR

This paper investigates how the effective depth of large language models varies with scale, training, and task difficulty, revealing underutilization of model layers and suggesting avenues for improving layer utilization.

Contribution

It systematically analyzes effective depth across model sizes and tasks, showing that current models underuse available layers regardless of scale or difficulty.

Findings

Effective depth ratio remains stable across model sizes.

Longer context does not increase effective depth.

Models do not use more layers for harder tasks.

Abstract

The scaling of large language models (LLMs) emphasizes increasing depth, yet performance gains diminish with added layers. Prior work introduces the concept of "effective depth", arguing that deeper models fail to fully utilize their layers for meaningful computation. Building on this, we systematically study how effective depth varies with model scale, training type, and task difficulty. First, we analyze the model behavior of Qwen-2.5 family (1.5B-32B) and find that while the number of effective layers grows with model size, the effective depth ratio remains stable. Besides, comparisons between base and corresponding long-CoT models show no increase in effective depth, suggesting that improved reasoning stems from longer context rather than deeper per-token computation. Furthermore, evaluations across tasks of varying difficulty indicate that models do not dynamically use more layers for harder problems. Our results suggest that current LLMs underuse available depth across scales, training paradigms and tasks of varying difficulties, pointing out research opportunities on increasing the layer utilization rate of LLMs, model pruning, and early exiting. Our code is released at https://github.com/AheadOFpotato/what_affects_effective_depth.