白泽(Whitzard)女娲实验室 ↗
NUWANUWA INDEX

搜索 NUWA Index

研究 · 测量协议 · 评测结果 · 观点

English
← 研究目录

风险评测 · 技术报告

GLM-5.2 Risk Evaluation Report

NUWA 研究解读 · 独立复核待完成

首次发表: 来源核验: 2026-10-25
NUWA Research Note研究问题研究设计与假设来源发现NUWA 解读局限与相反证据研究与决策意义机构与角色原始资源

模型辅助整理 · 选读原始全文或官方文章中的相关章节;尚未完成逐章节审计。

研究速读

NUWA 解读:推理预算能显著改变能力估计,防护充分性仍是另一问题。

研究问题

公共访问下独立GLM-5.2评估揭示哪些系统风险证据?

研究设计与假设

SaferAI未经开发者合作,从公开API测四类风险,采用公开基准和明确预算变化。

Agentic misalignment — Source figure
Source figure · 原文 Source figure。请结合下方的协议范围与局限阅读。 原图与来源 ↗

来源发现

无额外墙钟限制时,CyberGym复现从2M令牌36.6%升至50M的76.2%。

报告明确不据基准子集作整体风险判断。

NUWA 解读

NUWA 解读:推理预算能显著改变能力估计,防护充分性仍是另一问题。

局限与相反证据

公共访问与基准选择限制结论,跨领域得分不能相加为总风险。

研究与决策意义

模型建议:与系统卡比较前保留预算、访问方式和各基准防护条件。

精确来源位置
  • GLM-5.2 Risk Evaluation Report · Executive SummarySelected original resource https://www.safer-ai.org/research/glm-5-2-evaluation-report; sha256:c75f65dbb7b6a837496d0e97afe50d88948a0a878aa6d870c05493439f5ba8bd; archived source capture; exact capture time not independently established
  • GLM-5.2 Risk Evaluation Report · Cyber OffenseSelected original resource https://www.safer-ai.org/research/glm-5-2-evaluation-report; sha256:c75f65dbb7b6a837496d0e97afe50d88948a0a878aa6d870c05493439f5ba8bd; archived source capture; exact capture time not independently established
  • GLM-5.2 Risk Evaluation Report · ConclusionSelected original resource https://www.safer-ai.org/research/glm-5-2-evaluation-report; sha256:c75f65dbb7b6a837496d0e97afe50d88948a0a878aa6d870c05493439f5ba8bd; archived source capture; exact capture time not independently established
  • GLM-5.2 Risk Evaluation Report · Limitations and Future WorkSelected original resource https://www.safer-ai.org/research/glm-5-2-evaluation-report; sha256:c75f65dbb7b6a837496d0e97afe50d88948a0a878aa6d870c05493439f5ba8bd; archived source capture; exact capture time not independently established

来源核查、数据提取、编辑批准与实验复现分别记录。 方法说明 →

机构与角色

评测协议

Cybench →

CyberGym →

LAB-Bench →

BioMysteryBench →

SWE-Bench Pro →

Agentic misalignment →

MASK →

Attempt to Persuade Eval (APE) →

Petri →

引用与复用

引用研究发现时请引用原始出版物;引用 NUWA 的整理与解释时,单独引用本站记录及对应快照。

原始出版物 · BibTeX
@misc{saferai-glm-risk,
  title = {GLM-5.2 Risk Evaluation Report},
  url = {https://www.safer-ai.org/research/glm-5-2-evaluation-report},
  year = {2026},
  date = {2026-08-02},
  version = {Source capture 2026-10-25},
  note = {Source capture 2026-10-25}
}
NUWA 记录 · BibTeX
@misc{saferai-glm-risk-nuwa,
  title = {GLM-5.2 Risk Evaluation Report — NUWA research record},
  url = {https://whitzard.tech/nuwa/whitzard-index/work/saferai-glm-risk/},
  version = {2026-10-29.1},
  organization = {NUWA Frontier AI Safety Lab},
  howpublished = {Retained note edition: https://whitzard.tech/nuwa/whitzard-index/data/record-editions/2026-10-29.1--saferai-glm-risk.json},
  note = {NUWA research note. Source-reading scope and independent review status are stated on the record. Snapshot 2026-10-29.1. Retained note edition: https://whitzard.tech/nuwa/whitzard-index/data/record-editions/2026-10-29.1--saferai-glm-risk.json.}
}

证据快照 2026-10-29.1 · 索引更新 · RSS

参考文献

  1. [1] GLM-5.2 Risk Evaluation Report. 2026-08-02 ; Executive Summary; Cyber Offense; Conclusion
  2. [2] GLM-5.2 Risk Evaluation Report. 2026-08-02. v1.1 · revised 26 August 2026 ; Report masthead and stated evaluation methodology