When Does Multi-Agent RL Improve LLM Workflows?
Workflow, Scale, and Policy-Sharing Tradeoffs
Oregon State University · Pennsylvania State University · Adobe Inc. · AG2AI, Inc.
Abstract
Figure, table and section numbers in quoted paper text refer to the paper.
Setup
Three workflows, two policy-sharing strategies
Loading figures…
Takeaway
What multi-agent RL needs that single-agent RL does not
Citation
BibTeX