Do Agents Think Deeper? A Mechanistic Investigation of Layer-Wise Dynamics in Sequential Planning
This work provides mechanistic insights into how LLMs allocate depth adaptively in multi-turn planning, tool use, and iterative state updates, which is important for understanding and improving autonomous agent systems.
The paper investigates whether large language models use their depth more efficiently in autonomous agent settings compared to single-turn tasks. Through layer-wise analysis of agent trajectories, they find that agentic reasoning recruits deeper layers progressively, with a shift from feature accumulation to recalibration, and a construction-refinement gap where semantic direction forms early but deep layers are needed for stabilization.
Recent mechanistic studies suggest that large language models (LLMs) may utilize their depth inefficiently in standard single-turn tasks. Whether this still holds in autonomous agent settings, where models must perform multi-turn planning, tool use, and iterative state updates, remains unclear. We study this question through a systematic layer-wise analysis of complete user-agent trajectories spanning three domains: Deep Research, Code Generation, and Tabular Processing. Using residual stream probes, causal layer-skipping interventions, and effective-depth measurements, we show that agentic reasoning exhibits a distinct depth profile from static tasks. As trajectories unfold, models progressively recruit more and deeper layers, with stronger long-range inter-layer dependencies emerging in later turns. At the same time, residual updates become increasingly correction-dominant, indicating a shift from stable feature accumulation toward repeated recalibration. Effective-depth analysis further reveals a substantial construction-refinement gap: semantic direction often forms relatively early, while deep layers remain necessary for stabilizing final outputs. Across model families, this gap is pronounced in Qwen and Minimax, whereas GLM shows a more domain-dependent depth allocation pattern. These results provide mechanistic evidence that autonomous LLM agents allocate depth adaptively as reasoning complexity grows.