三种角色，一个模型：推理时的角色编排以缩小小规模与大规模代理之间的性能差距
Three Roles, One Model: Role Orchestration at Inference Time to Close the Performance Gap Between Small and Large Agents
AI总结：
本文研究在推理时仅使用架构支撑是否能提升小模型在复杂多步骤环境中的性能，通过三阶段推理架构使小模型在AppWorld基准测试中任务完成率提升近一倍，展示了结构化推理干预使小模型能与大模型竞争。
摘要：
大型语言模型（LLM）代理在现实工具使用任务中表现出色，但在有限硬件上部署有能力的代理仍具挑战性。我们研究在推理时仅使用架构支撑（不增加额外训练计算）是否能提升小模型在复杂多步骤环境中的性能。在单块24GB GPU上，我们评估Qwen3-8B在AppWorld基准测试下的全精度和4位量化配置。未经干预，原始模型仅能完成5.4%（FP16）和3.0%（AWQ）的任务目标。通过系统性失败模式分析，我们引入一个三层推理架构管道，部署相同的冻结模型在三个不同角色：（1）一个摘要模型，保留关键物品（令牌、凭证、API响应）同时压缩对话历史；（2）主代理模型，对压缩上下文进行推理；（3）一个隔离的修正模型，审查并修订代理的代码输出，不访问对话历史，打破重复失败循环。应用于相同未修改的模型，此架构使任务完成率提升至8.9%（FP16）和5.9%（AWQ），在两种设置中性能大致翻倍，尤其在难度1任务中（FP16：15.8%到26.3%；AWQ：5.3%到14.0%）。在全精度推理中，我们的架构8B模型超越了DeepSeek-Coder 33B Instruct（7.1%）的原始AppWorld评估结果，证明了结构化推理时干预可使小模型与四倍于自身的系统竞争。我们将其形式化为一个基于冻结基础模型的架构策略，三次不同条件下的相同权重调用，连接到测试时计算扩展和动作空间塑造在强化学习中的概念。
英文摘要：
Large language model (LLM) agents show promise on realistic tool-use tasks, but deploying capable agents on modest hardware remains challenging. We study whether inference-time scaffolding alone, without any additional training compute, can improve the performance of a small model in complex multi-step environments. Operating on a single 24GB GPU, we evaluate Qwen3-8B on the AppWorld benchmark under both full-precision and 4-bit quantized configurations. Without any intervention, the raw model achieves just 5.4% (FP16) and 3.0% (AWQ) task goal completion. Guided by a systematic failure mode analysis, we introduce a three-tier inference scaffolding pipeline that deploys the same frozen model in three distinct roles: (1) a summarization model that preserves critical artifacts (tokens, credentials, API responses) while compressing dialogue history; (2) the main agent model that reasons over the compressed context; and (3) an isolated correction model that reviews and revises the agent's code output without access to conversation history, breaking repetitive failure loops. Applied to the same unmodified model, this scaffolding yields 8.9% (FP16) and 5.9% (AWQ) task goal completion, roughly doubling performance in both settings, with particularly strong gains on difficulty-1 tasks (15.8% to 26.3% FP16; 5.3% to 14.0% AWQ). On full-precision inference, our scaffolded 8B model surpasses DeepSeek-Coder 33B Instruct (7.1%) from the original AppWorld evaluation, demonstrating that structured inference-time interventions can make small models competitive with systems 4 times their size. We formalize the approach as a scaffolded policy over a frozen base model, three invocations of the same weights with different conditioning, drawing connections to test-time compute scaling and action-space shaping in reinforcement learning.
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。
