在大规模工具空间中扩展代理能力，而非上下文：用于大规模工具空间的高效强化微调
Scaling Agentic Capabilities, Not Context: Efficient Reinforcement Finetuning for Large Toolspaces
AI总结：
ATLAS通过强化微调框架使小型语言模型在大规模工具空间中更高效地执行任务，通过学习上下文获取和动作执行，提升长周期工作流程的稳定性与效率。
摘要：
在大规模工具生态系统上运行的代理系统必须在弱监督或不可验证的监督下规划和执行长周期工作流程。虽然前沿模型通过规模和大上下文预算缓解这些挑战，但小型语言模型（SLMs）仍然脆弱：急于加载工具会耗尽上下文，执行错误会随时间累积，稀疏奖励限制学习。我们介绍了ATLAS，一种强化微调框架，使SLMs能够在大规模工具空间环境中有效运作，通过学习如何获取上下文和如何执行动作。我们的方法有两大关键贡献。首先，我们将上下文控制和执行结构视为可学习的决策，结合迭代工具加载与程序化工具编排，以限制上下文增长并稳定长周期轨迹。其次，我们提出了基于规则的强化微调，将任务成功分解为结构化、任务对齐的准则，并通过小型判断模型实现可扩展的训练。在MCP基准测试中，这些设计选择在通用RL基线上产生了显著且一致的提升，使400亿参数的SLM在远更紧的参数和上下文预算下接近前沿代理性能。
英文摘要：
Agentic systems operating over large tool ecosystems must plan and execute long-horizon workflows under weak or non-verifiable supervision. While frontier models mitigate these challenges through scale and large context budgets, small language models (SLMs) remain brittle: eager tool loading saturates context, execution errors compound over time, and sparse rewards limit learning. We introduce ATLAS, a reinforcement finetuning framework that enables SLMs to operate effectively in large-scale toolspace environments by learning how to acquire context and how to execute actions. Our approach makes two key contributions. First, we treat context control and execution structure as learnable decisions, combining iterative tool loading with programmatic tool orchestration to bound context growth and stabilize long-horizon trajectories. Second, we propose rubric-based reinforcement finetuning, which decomposes task success into structured, task-aligned criteria and enables scalable training using small judge models. Across MCP benchmarks, these design choices yield large and consistent gains over generic RL baselines, allowing a 4B SLM to approach frontier-agent performance under far tighter parameter and context budgets.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
