从单打独斗到社会学习：LLM 递归社会性改进能力研究

View PDF HTML (experimental)

Abstract:Large language models (LLMs) can now improve themselves by revising the instructions they follow, and LLM agents are increasingly orchestrated to work together on complex problems. However, self-improvement methods typically optimize one system at a time, and multi-agent frameworks often have every model work toward a shared goal. We ask a different question. When each agent pursues its own reward, can self-improving LLMs learn from one another well enough to improve the whole population? We call this capability recursive social improvement. We study populations that revise skill files and choose whether, when, and whom to copy from. Independent search, learning from peers, and acting all share one token budget. In controlled environments, established social-learning algorithms benefit from peers, but three LLMs do not. They earn less reward per token than solo learners, and explore too narrowly or run out of tokens before acting. We then let the models write and revise their own skills. Observing peers changes how they improve, helping one model find useful skills sooner and another spend less on private search. Neither, however, outperforms independent learners at the same cost. Skills are copied, revised, and passed on, so one discovery can seed further search. Yet these exchanges concentrate the population around fewer independent discoveries. Together, these results show that LLMs can make learning more efficient by copying from peers, but not yet more effective.

Subjects:	Multiagent Systems (cs.MA); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Cite as:	arXiv:2609.38516 [cs.MA]
	(or arXiv:2609.38516v2 [cs.MA] for this version)
	https://doi.org/10.48550/arXiv.2609.38516 arXiv-issued DOI via DataCite

Submission history

From: Kunal Jha [view email]
[v1] Tue, 29 Sep 2026 20:37:10 UTC (1,393 KB)
[v2] Thu, 1 Oct 2026 21:46:00 UTC (1,393 KB)