ACL Findings2026
AJ-Bench: Benchmarking Agent-as-a-Judge for Environment-Aware Evaluation
Wentao Shi, Yu Wang, Yuyang Zhao, Yuxin Chen, Fuli Feng, Xueyuan Hao, Xi Su, Qi Gu, Hui Su, Xunliang Cai, Xiangnan He
摘要
As reinforcement learning continues to scale the training of large language model-based agents, reliably verifying agent behaviors in complex environments has become increasingly challenging. Existing approaches rely on rule-based verifiers or LLM-as-a-Judge models, which struggle to generalize beyond narrow domains. Agent-as-a-Judge addresses this limitation by actively interacting with environments and tools to acquire verifiable evidence, yet its capabilities remain underexplored.
We introduce a benchmark AJ-Bench to systematically evaluate Agent-as-a-Judge across three domains-search, data systems, and graphical user interfaces-comprising 155 tasks and 516 annotated trajectories. The benchmark comprehensively assesses judge agents' abilities in information acquisition, state verification, and process verification. Experiments demonstrate consistent performance gains over LLM-as-a-Judge baselines, while also revealing substantial open challenges in agent-based verification. Our data and code are available at https://aj-bench.github.io/.
问 Lune
几分钟读透这篇 Paper,不用再花一下午。
Lune 能读完全文,也能读引用它的顶会 Paper,讲清它的结论、证据和局限,每一点都注明出自哪篇 Paper。不用硬啃,也不靠猜。
脉络
它引用的 Paper20
- WebArena: A Realistic Web Environment for Building Autonomous AgentsShuyan Zhou, Frank F. Xu, Hao Zhu, Xuhui Zhou 等ICLR 2024 · 1,197 次引用
- Evaluating Large Language Models at Evaluating Instruction FollowingZhiyuan Zeng, Jiatong Yu, Tianyu Gao, Yu Meng 等ICLR 2024 · 299 次引用
- MetaTool Benchmark for Large Language Models: Deciding Whether to Use Tools and Which to UseYue Huang, Jiawen Shi, Yuan Li, Chenrui Fan 等ICLR 2024 · 188 次引用
- Reinforcement Learning with Verifiable Rewards: GRPO's Loss, Dynamics, and Success AmplificationYoussef MrouehICML 2026 · 118 次引用
- API-Bank: A Comprehensive Benchmark for Tool-Augmented LLMsMinghao Li, Yingxiu Zhao, Bowen Yu, Feifan Song 等EMNLP 2023 · 72 次引用
相关 Paper
- VerifyBench: A Systematic Benchmark for Evaluating Reasoning Verifiers Across DomainsXuzhao Li, Xuchen Li, Shiyu Hu, Yongzhen Guo 等AAAI 2026 · 16 次引用
- MAS-ProVe: Understanding the Process Verification of Multi-Agent SystemsVishal Venkataramani, Haizhou Shi, Zixuan Ke, Austin Xu 等ICML 2026 · 5 次引用
- AssistantBench: Can Web Agents Solve Realistic and Time-Consuming Tasks?Ori Yoran, Samuel Joseph Amouyal, Chaitanya Malaviya, Ben Bogin 等EMNLP 2024 · 5 次引用
- ProRe: A Proactive Reward System for GUI Agents via Reasoner-Actor CollaborationGaole Dai, Shiqi Jiang, Ting Cao, Yuqing Yang 等ICLR 2026 · 10 次引用
- A Benchmark for Deep Information SynthesisDebjit Paul, Daniel Murphy, Milan Gritta, Ronald Cardenas 等ICLR 2026 · 1 次引用
