RLCR trains LLM agents as organizations: emergent hierarchies and division of labour
algo_diver · x · 2026-10-24
A new paper introduces Reinforcement Learning from Collective Rewards (RLCR), a recipe that trains language models as agent organizations sharing a message board and workspace, used to train Avalon-M1-Mini (120B).
Key findings:
- Agents spontaneously learn organizational hierarchies and division of labour
- Increased collective verification, including hiring verifiers and testers
- The rise of collective thinking, where agents reason about helping other agents
The authors position RLCR as a laboratory for studying collective intelligence and multi-agent safety. Lewis Tunstall predicts the paper is about to get "1000x more citations."
More from Research
- FADA adapts humanoid robots to new conditions in ~2 minutes, no rewards needed — GuanyaShi · 2026-10-24
- Energy-Based Transformer research teased, with claims diffusion's end is near — burny_tech · 2026-10-24
- "Looped Model Atlas" visualizes LLM architecture papers as an interactive magic circle — burny_tech · 2026-10-24
- Rocq formalization shows Peano Arithmetic proves "K230 halts iff NOT Con(ZFC)" — burny_tech · 2026-10-24
- AI now designs custom phages on demand to kill superbugs, replacing sewage hunting — RobertWKemp · 2026-10-24
- Reviewer chimes in: 'I once reviewed a paper with a negative R²' — RexDouglass · 2026-10-24