跳到正文
HuggingFace Daily Papers(社區熱門論文)·· 2026-10-12AI 評分49

Recursive Environment Scaling for General Long-Horizon Agents

Recursive Environment Scaling for General Long-Horizon Agents

AI 導讀

Recursive Environment Scaling(RES)透過遞迴改造基準環境,在保留核心任務目標的同時提高難度,用於評估 Agent 及收集微調軌跡。在 294 項任務的實驗中,六個評測模型的任務成功率均下降;以 RES 軌跡微調,九組模型與基準的表現均較原始軌跡 SFT 高 7.00–14.61 個百分點。

正文

Published on Oct 12

Authors:

,

,

,

,

,

,

,

,

,

,

,

Abstract

Recursive Environment Scaling transforms benchmarks to increase difficulty while preserving objectives, improving both agent evaluation and supervised fine-tuning performance.

We study environment scaling as a way to increase agent-benchmark difficulty while holding core task objectives fixed. Our framework, Recursive Environment Scaling (RES), transforms existing benchmark environments while preserving their objectives and completion requirements. A proposer changes information access, resource relations, and execution dependencies; execution-grounded checks validate the resulting candidate, and accepted changes are inherited across rounds. This produces related environments for paired evaluation and training-trajectory collection. Experiments on 294 initial tasks from three benchmarks show lower task success for all six evaluation models, with decreases of 57.14--77.14 percentage points on Claw-Eval-Live. With equal trajectory counts and identical optimization settings, fine-tuning on RES trajectories outperforms original-trajectory SFT in all nine model--benchmark comparisons by 7.00--14.61 percentage points. Ablations identify recursive inheritance and dependency composition as important contributors to increased difficulty. Together, these results support environment scaling for both agent evaluation and training.

View arXiv page View PDF Add to collection

Get this paper in your agent:

hf papers read 2610.16397

Don't have the latest CLI?

curl -LsSf https://hf.co/cli/install.sh | bash

Models citing this paper 0

No model linking this paper

Cite arxiv.org/abs/2610.16397 in a model README.md to link it from this page.

Datasets citing this paper 0

No dataset linking this paper

Cite arxiv.org/abs/2610.16397 in a dataset README.md to link it from this page.

Spaces citing this paper 0

No Space linking this paper

Cite arxiv.org/abs/2610.16397 in a Space README.md to link it from this page.

Collections including this paper 0

No Collection including this paper

Add this paper to a collection to link it from this page.

來源:HuggingFace Daily Papers(社區熱門論文) · huggingface.co