Saved in:
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2602.03224 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866912871821082624 |
|---|---|
| author | Cheng, Yu Zhou, Jiuan Hu, Yongkang Chen, Yihang Zhou, Huichi Chen, Mingang Zhang, Zhizhong Shao, Kun Xie, Yuan Yin, Zhaoxia |
| author_facet | Cheng, Yu Zhou, Jiuan Hu, Yongkang Chen, Yihang Zhou, Huichi Chen, Mingang Zhang, Zhizhong Shao, Kun Xie, Yuan Yin, Zhaoxia |
| contents | Test-time evolution of agent memory serves as a pivotal paradigm for achieving AGI by bolstering complex reasoning through experience accumulation. However, even during benign task evolution, agent safety alignment remains vulnerable-a phenomenon known as Agent Memory Misevolution. To evaluate this phenomenon, we construct the Trust-Memevo benchmark to assess multi-dimensional trustworthiness during benign task evolution, revealing an overall decline in trustworthiness across various task domains and evaluation settings. To address this issue, we propose TAME, a dual-memory evolutionary framework that separately evolves executor memory to improve task performance by distilling generalizable methodologies, and evaluator memory to refine assessments of both safety and task utility based on historical feedback. Through a closed loop of memory filtering, draft generation, trustworthy refinement, execution, and dual-track memory updating, TAME preserves trustworthiness without sacrificing utility. Experiments demonstrate that TAME mitigates misevolution, achieving a joint improvement in both trustworthiness and task performance. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_03224 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | TAME: A Trustworthy Test-Time Evolution of Agent Memory with Systematic Benchmarking Cheng, Yu Zhou, Jiuan Hu, Yongkang Chen, Yihang Zhou, Huichi Chen, Mingang Zhang, Zhizhong Shao, Kun Xie, Yuan Yin, Zhaoxia Artificial Intelligence Machine Learning Test-time evolution of agent memory serves as a pivotal paradigm for achieving AGI by bolstering complex reasoning through experience accumulation. However, even during benign task evolution, agent safety alignment remains vulnerable-a phenomenon known as Agent Memory Misevolution. To evaluate this phenomenon, we construct the Trust-Memevo benchmark to assess multi-dimensional trustworthiness during benign task evolution, revealing an overall decline in trustworthiness across various task domains and evaluation settings. To address this issue, we propose TAME, a dual-memory evolutionary framework that separately evolves executor memory to improve task performance by distilling generalizable methodologies, and evaluator memory to refine assessments of both safety and task utility based on historical feedback. Through a closed loop of memory filtering, draft generation, trustworthy refinement, execution, and dual-track memory updating, TAME preserves trustworthiness without sacrificing utility. Experiments demonstrate that TAME mitigates misevolution, achieving a joint improvement in both trustworthiness and task performance. |
| title | TAME: A Trustworthy Test-Time Evolution of Agent Memory with Systematic Benchmarking |
| topic | Artificial Intelligence Machine Learning |
| url | https://arxiv.org/abs/2602.03224 |