跳到正文
HuggingFace Daily Papers(社區熱門論文)·· 2026-10-11AI 評分50

Scalable Weak-to-Strong Verifier Hardening for Agent Environments

Scalable Weak-to-Strong Verifier Hardening for Agent Environments

AI 導讀

研究提出一套可擴展、全自動的弱至強驗證器強化方法,透過隔離沙盒及「攻擊者—修復者」迴圈,反覆找出並修補智能體環境驗證器的漏洞,以防止獎勵作弊。在 Terminal Bench、KernelBench 和 SETA 上,方法大幅降低攻擊成功率,且迭代次數增加時防禦能力逐步提升,表現優於非迭代基線。較早版本的迴圈已獲 Terminal Bench 官方存放庫採用,用於審核環境中的漏洞利用。

正文

Published on Oct 11

Authors:

,

,

,

,

,

,

Abstract

The study proposes an automated weak-to-strong loop that hardens agent environment verifiers against reward hacking by iteratively finding and repairing vulnerabilities in isolated sandboxes.

Agent environments are essential for large language model (LLM) development. They use programmatic verifiers to turn complex outcomes into rewards. Verifiers can be attacked into awarding full credit to cheating agents without completing the intended task, causing inaccurate evaluation and misalignment in training. As agent environments scale in numbers and difficulty, so must the repairing of exploitable verifiers. To this end, we ask whether today's model with scaled compute can automatically harden agent environments to prevent potentially stronger future models from reward hacking. We introduce a scalable, fully automatic approach to weak-to-strong verifier hardening. It first sets up an isolated verifier sandbox that agents cannot directly modify, then uses a hacker-fixer loop to find and repair remaining verifier vulnerabilities. We apply this loop to Terminal Bench, KernelBench, and SETA, substantially reducing attack success across these agent environments. The iterative design scales with the hardening budget, producing progressively stronger defenses as the number of loop iterations increases and outperforming the static, non-iterative baseline. An earlier version of our hacker-fixer loop has been adopted by the official Terminal Bench repository as a way to audit exploits in its environments. Project page: https://verifier-hardening.github.io/.

View arXiv page View PDF Project page Add to collection

Get this paper in your agent:

hf papers read 2610.15924

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/2610.15924 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/2610.15924 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/2610.15924 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