#stress testing

5 stories · 0 in 7 days · 1 in 30 days

AI stocks 1 BOJ 1 PACT 1 enterprise AI 1

Stories, products and papers tagged "stress testing", newest first.

Oct 24

general Policy01:04

Chinese content日本央行压力测试：银行已准备好应对 AI 股崩盘与全球利率飙升

日本央行测了下银行扛不扛得住 AI 股崩盘加全球加息，结果是有惊无险，做风控的可以看看情景怎么设的。

#BOJ #stress testing #AI stocks

Nikkei Asia@NikkeiAsiaSource

Sep 17

Papers BuildersSelected09:27

PACT benchmark for AI assistants' compliance under pressure

This research provides a new way to measure how reliable AI assistants are in sensitive business contexts, which is important for companies considering deploying such tools.

#PACT #enterprise AI #stress testing

arXiv cs.AI@Mika Okamoto, Ansel Kaplan ErolSource

Aug 10

Papers11:11

Chinese contentSABRE：可扩展自动化的VLM压力测试基准构建流程

这论文给了一个能自动造VLM压力测试的框架，比现有基准更能暴露模型短板，做评测的可以看看。

#SABRE #VLM #stress testing #benchmarks

arXiv cs.AI@Zixuan Lan et al. (4)Source

Jun 29

Papers US vs ChinaSelected15:42

Chinese contentNature Medicine论文开源测试套件：GPT-5.5 Pro得分79%

Nature Medicine那篇论文把模型考倒了，但作者直接把考卷开源了。后来GPT-5.5 Pro重新考，分数涨了10%！

#GPT-5 #Gemini 2.5 Pro #stress testing #open source

@koltregaskes@koltregaskesSource

Jun 11

Papers11:49

Chinese contentLLM 安全裁判在翻译后改变判决，论文揭示一致性危机

做 AI 安全评测的团队会直接受影响——你的安全裁判可能比想象中更不可靠，建议点开看看测试方法。

#LLM safety #consistency #stress testing #arxiv

rohanpaul_ai@rohanpaul_aiSource