Tag
研究者构建了基于 Pokémon Showdown 2v2 的多智能体环境，评估 LLM 代理的协作能力。结果显示自由对话模式胜率达 64%，显著高于无对话的 42%，表明对话本身是协作收益的关键来源。
该论文研究多智能体 LLM 谈判系统中的三部分控制栈：5 条支柱运行时宪章、4 层 swarm 架构（Director、多数投票、Monitor、schema 硬门控）以及 Cognitive Annealing（确定性死锁检测与原子上下文清除恢复），并针对一个有状态对抗性 Gatekeeper 进行消融实验，发现确定性恢复机制在死锁场景中显著优于纯 LLM 引导。
This paper introduces an exact generate–transform decomposition to analyze how small 7–9B LLM teams scale across eight orchestration architectures, finding that scaling returns are sharply task-dependent — Proposer-Critic wins on arithmetic benchmarks but no architecture dominates across all tasks.
The paper introduces Learning What to Skip (LW2S), a method that uses counterfactual credit assignment to optimize multi-agent LLM workflows by selectively skipping components, reducing token cost while maintaining or improving accuracy.
TOPAS is a workflow-aware prefix-state scheduling method for multi-agent LLM serving that optimizes cache usage and request scheduling to reduce job completion time.
The developer of GenOS, an open-source framework for multi-agent LLM orchestration using Rust and Git worktrees, is requesting community feedback with worst edge cases to stress-test and improve the system.