跳到正文
arXiv cs.AI· Rui Li, Liyang He, Zheng Zhang, Zhenya Huang, Qi Liu·· 10 小时前AI 评分48

AnyS1:为任意 Agent 工作流决策点提供统一的 System-1 模型

Thinking Fast, Anywhere: One System-1 for the Decision Points of Any Agent Workflow

AI 导读

AnyS1 框架将人类编写的流程技能转化为跨领域通用的决策点来源,并训练单个 4B 指针 System-1 模型,在运行时一次前向传播处理常规决策、置信度低于校准阈值时回退 System-2。在三个未见基准和不同 System-2 模型族上,AnyS1 保持任务可靠性,同时将单次任务成本降低 75-91%、模型调用时间减少 63-92%,并支持零样本阈值迁移。

正文 · 原文

View PDF HTML (experimental)

Abstract:Autonomous LLM agents incur substantial cost and latency by invoking a large model (System-2) at every step, even when many intermediate actions choose among a small set of known candidates. Existing dual-system agents offload such decisions to a lightweight System-1 model, but rely on workflow-specific hooks authored per environment. We present AnyS1, a framework that turns human-authored procedural skills, whose instructions naturally specify where choices arise and which options apply, into a shared, domain-general source of decision points. Offline, AnyS1 mines grounded enumerable decisions across over 1,000 public skills, gathers execution-grounded trajectory states inside isolated containers to construct Any-DP, and trains a single calibrated 4B pointer System-1 with label smoothing. At run time, compiled workflows resolve routine decisions in one forward pass and fall back to System-2 when confidence falls below a calibrated threshold. Across three unseen benchmarks and distinct System-2 model families, AnyS1 preserves task reliability while cutting episode cost by 75-91% and model-call time by 63-92% relative to an uncompiled System-2 agent, and maintains performance under zero-shot threshold transfer.
Comments: 16 pages, 8 figures, 7 tables. Project page: this https URL
Subjects: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)
Cite as: arXiv:2611.23751 [cs.AI]
  (or arXiv:2611.23751v1 [cs.AI] for this version)
  https://doi.org/10.48550/arXiv.2611.23751

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Rui Li [view email]
[v1] Wed, 18 Nov 2026 03:05:41 UTC (203 KB)

来源:arXiv cs.AI · arxiv.org