ExRole: From Team Trajectories to Executable Roles in Multi-Agent Language Models

2026-08-12

AI Summary

This work addresses the limitations of existing role modeling approaches, which predominantly rely on static prompt tags and struggle to capture learning behaviors or guide trainable capabilities. The paper proposes ExRole, a novel framework that redefines roles as executable control variables, enabling the learning of future-oriented role prototypes from team interaction trajectories. ExRole jointly generates human-readable instructions and aligned token markers, unifying behavior prediction, interaction guidance, and capability localization. The framework introduces trajectory-based role learning and a turn-level credit assignment mechanism leveraging shared LoRA rank slots, facilitating dynamic role routing across agents and turns. Experimental results demonstrate that ExRole substantially outperforms single-agent baselines by 15.0/14.4 EM/F1 points on MuSiQue and 13.5/16.1 EM/F1 points on 2WikiMultiHopQA, significantly exceeding various ablation and comparison settings.

Abstract

Roles provide an interpretable interface for organizing language-model agents, yet most multi-agent systems treat them as hand-written prompt labels disconnected from learned behavior and parameter updates. We argue that a useful role should instead be an executable control variable: it should summarize behavior predictive of future utility, guide subsequent interaction, and identify the trainable capacity responsible for that behavior. We introduce ExRole, a trajectory-to-role framework that learns future-aware role prototypes from prefix-local team traces, resolves them into readable instructions and token-aligned role markers, and optionally routes shared LoRA rank slots with turn-aligned credit. Across MuSiQue and 2WikiMultiHopQA, ExRole improves over single-agent search by 15.0/14.4 and 13.5/16.1 EM/F1 points, respectively. Against the strongest non-ExRole controls, the corresponding gains remain 11.5/11.6 and 7.7/9.7 points. Across both benchmarks, the controlled results consistently favor trajectory-induced role conditioning over role-free, manual, random, and shuffled alternatives. Role-Agent-Turn interventions further show that the induced roles capture transferable behavioral specialization beyond fixed agent identities or turn positions.

Problem

• multi-agent language models
• executable roles
• role conditioning
• behavioral specialization
• trajectory-based learning

Innovation

• executable roles
• trajectory-to-role learning
• LoRA routing
• multi-agent language models
• behavioral specialization