arXiv:2609.23989 · 工程化
ACLArena: Agent Continue Learning in Multi-stage Post-training
ACLArena：多阶段后训练中的 Agent 持续学习
ACLArena: Agent Continue Learning in Multi-stage Post-training
- 类型：arxiv
- 标识：2609.23989
- 链接：https://arxiv.org/abs/2609.23989
- 主分类：engineering
- 形态：application
- TLDR：Building general-purpose agents for industrial deployment requires integrating multiple capabilities, each typically acquired at a distinct stage of training. Yet there is currently no well-established recipe for Agent Continual Learning (ACL), with little understanding of the trade-offs among existing integration paradigms. To address this gap, we introduce ACLArena, a framework for comprehensively studying, analyzing, and evaluating ACL. We first build a sequential training pipeline and conduct an in-depth analysis that explains the mechanisms of forgetting and generalization from two comple
- 副分类：agent
- 待LLM分类：否
- 标题中文：ACLArena：多阶段后训练中的 Agent 持续学习
- TLDR中文：构建面向工业部署的通用 agent 需要整合多种能力，而每种能力通常在不同的训练阶段获得。然而目前尚无成熟的 Agent 持续学习（ACL）方案，对现有整合范式之间的权衡也缺乏理解。为填补这一空白，我们提出 ACLArena，一个用于全面研究、分析和评估 ACL 的框架。我们首先构建了一条顺序训练 pipeline，并从两个互补的视角深入分析了遗忘与泛化的机制。
- 来源文件：
- /inbox/tom/_candidates/2026-09-23-agent-rag-longcontext-candidates.json