cs.CL提交于 2026-10-26已译

超越最终结果：量化多智能体 LLM 交互中的声明级信息动态

Beyond Final Outcomes: Quantifying Claim-Level Information Dynamics in Multi-Agent LLM Interaction

Haocheng Jiang · Herun Wan · Yongyuan Peng · Minnan Luo

中文摘要

当基于 LLM 的智能体进行交互时，信息会不断被共享、扩展和遗忘，但多智能体评估通常仅关注最终任务结果。该研究将交互视为声明级信息动态，并提出 RIVER：将智能体输出分解为原子声明，评估收敛性、探索性和保留性三个互补维度。收敛性衡量跨智能体声明的对应关系与立场一致性；探索性衡量语义广度及对来源内容的扩展；保留性追踪来源信息的保存情况，以及新声明在阶段和谱系间的持续性。在 40 个受控新闻生成案例中评估了 5 种交互策略。反复汇集式交互能促进收敛，但会缩窄信息空间；分布式接力则产生更广泛、共享程度更低的信息。进一步比较表明，方法特定约束可提高探索性，同时削弱收敛性和保留性。研究结果显示，相比下游表现，底层信息流机制更能解释集体信息动态。

关键要点

01多智能体评估通常只看最终任务结果，难以刻画信息在交互中的共享、扩展与遗忘。
02RIVER 将输出分解为原子声明，从收敛性、探索性和保留性三个维度量化信息动态。
0340 个受控新闻生成案例比较了 5 种交互策略，揭示不同信息流机制的差异。
04反复汇集式交互促进收敛但缩窄信息空间，分布式接力带来更广泛且共享更少的信息。
05方法特定约束会增强探索性，同时削弱收敛性与保留性。

解读

尚无解读。

原始英文摘要

arXiv:2610.32333v1 Announce Type: new Abstract: When LLM-powered agents interact, information is repeatedly shared, expanded, and forgotten, yet multi-agent evaluation typically focuses on final task outcomes. We study such interaction as claim-level information dynamics and propose RIVER, which decomposes agent outputs into atomic claims and evaluates three complementary dimensions: convergence, exploration, and retention. Convergence captures cross-agent claim correspondence and stance alignment; exploration captures semantic breadth and extension beyond the source; and retention tracks source preservation and the persistence of new claims across stages and lineages. We evaluate five interaction strategies on 40 controlled news-generation cases. Repeatedly pooled interaction promotes convergence but narrows the information space, whereas distributed relay yields broader, less shared information. Further comparisons show that method-specific constraints increase exploration while weakening convergence and retention. Our findings suggest that collective information dynamics are better explained by underlying information-flow mechanisms than by simple downstream performance.