arXiv Machine Learning July 27, 2026 By Xiaoyang Cao, Siddarth Srinivasan, Michiel A. Bakker

Do Modules Stay in Their Lane? Role Drift in Compound LLM Systems

Read the original on arXiv Machine Learning →

arXiv:2607. 21627v1 Announce Type: cross Abstract: End-to-end reinforcement learning can improve the accuracy of compound LLM systems, but it does not constrain how modules divide labor internally.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Machine Learning.

arXiv AI

Jun 2

When Does Multi-Agent RL Improve LLM Workflows? Workflow, Scale, and Policy-Sharing Tradeoffs

arXiv:2605. 24202v2 Announce Type: replace Abstract: Multi-agent LLM workflows route inference through specialized roles to lift end-task accuracy, but jointly training those roles with reinforcement learning is unstable in ways that are poorly understood.

By Yifan Zeng, Yiran Wu, Yaolun Zhang, Wentian Zhao, Kun Wan, Qingyun Wu, Huazheng Wang

llms agents reinforcement-learning

arXiv AI

Jul 20

PASs-MoE: Mitigating Misaligned Co-drift among Router and Experts via Pathway Activation Subspaces for Continual Learning

arXiv:2601. 13020v2 Announce Type: replace-cross Abstract: Continual instruction tuning (CIT) requires multimodal large language models (MLLMs) to adapt to a stream of tasks without forgetting prior capabilities.

By Zhiyan Hou, Haiyun Guo, Haokai Ma, Yandu Sun, Yonghui Yang, Jinqiao Wang

llms fine-tuning multimodal benchmarks

arXiv Machine Learning

Sep 15

What Does an LLM Learn from Reinforcement Learning? A Mechanistic Interpretability Perspective with Fixed-SAE Track

arXiv:2609.15064v1 Announce Type: new Abstract: Reinforcement learning (RL) is widely utilized in large language model training to improve targeted capabilities, yet how RL reshapes a model remains p...

By Lingheng Du, Yiming Tang, Xufeng Duan, Dianbo Liu

llms reinforcement-learning benchmarks safety

arXiv AI

Sep 1

Drop the Act: Probe-Filtered RL for Faithful Chain-of-Thought Reasoning

arXiv:2605.11467v2 Announce Type: replace-cross Abstract: Reasoning models post-hoc rationalize answers they have already committed to internally, producing chains of *reasoning theater*: deliberativ...

By Swapnil Parekh, Naman Goyal

llms reinforcement-learning benchmarks safety

arXiv AI

Sep 15

The Router Within: Eliciting Native Skill Routing from a Frozen LLM

arXiv:2609.15982v1 Announce Type: cross Abstract: Skills extend an LLM agent beyond its parametric knowledge, and the gain they promise rests on picking the right one. Deployed harnesses route by pre...

By Ruishuo Chen, Xun Wang, Yu Chen, Zhuoran Li, Longbo Huang

llms agents benchmarks

arXiv AI

Jul 9

RL Post-Training Builds Compositional Reasoning Strategies

arXiv:2607. 07646v1 Announce Type: new Abstract: Does RL post-training merely amplify primitive skills already latent in a base model, or can it compose primitive skills into new higher-level strategies?

By Azwar Abdulsalam, Nishil Patel, Andrew Saxe

llms reinforcement-learning fine-tuning

One email a morning, machine-written

Related stories

When Does Multi-Agent RL Improve LLM Workflows? Workflow, Scale, and Policy-Sharing Tradeoffs

PASs-MoE: Mitigating Misaligned Co-drift among Router and Experts via Pathway Activation Subspaces for Continual Learning

What Does an LLM Learn from Reinforcement Learning? A Mechanistic Interpretability Perspective with Fixed-SAE Track

Drop the Act: Probe-Filtered RL for Faithful Chain-of-Thought Reasoning

The Router Within: Eliciting Native Skill Routing from a Frozen LLM

RL Post-Training Builds Compositional Reasoning Strategies