PulseAugur
中
实时 12:14:21
English(EN) A stronger model had a weaker lock UK AISI broke pre-release GPT-6.2 Astra with half the queries needed for GPT-6 Astra. OpenAI’s automated testing found two hi

英国人工智能安全研究所发现 GPT-6.2 Astra 比 GPT-6 Astra 更容易被越狱

据报道,英国人工智能安全研究所 (AISI) 发现,OpenAI 的 GPT-6.2 Astra 的预发布版本比其前身 GPT-6 Astra 更容易被越狱。与前一个模型相比,AISI 的测试仅用了一半的查询就实现了越狱。然而,OpenAI 自有的自动化测试发现了两个高危越狱,而四个私人红队则未发现。 AI

影响 凸显了人工智能模型的潜在安全问题以及不同测试方法有效性的差异。

排序理由 该条目讨论了人工智能安全研究所关于模型漏洞的发现,属于人工智能安全研究的范畴。[lever_c_demoted from research: ic=1 ai=1.0]

在 Mastodon — mastodon.social 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。 我们如何撰写摘要 →

英国人工智能安全研究所发现 GPT-6.2 Astra 比 GPT-6 Astra 更容易被越狱

本文如何被排名

Signal score
19 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
该条目讨论了人工智能安全研究所关于模型漏洞的发现,属于人工智能安全研究的范畴。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration
Single-source cluster
Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics
safety, model release
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
Breaking (< 6h)
Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

完整方法见我们的编辑标准。

报道来源 [1]

  1. Mastodon — mastodon.social TIER_1 English(EN) · spaisee ·

    更强的模型拥有更弱的锁 UK AISI 在 GPT-6.2 Astra 发布前破解,仅用 GPT-6 Astra 一半的查询次数。OpenAI 的自动化测试发现了两个 hi

    A stronger model had a weaker lock UK AISI broke pre-release GPT-6.2 Astra with half the queries needed for GPT-6 Astra. OpenAI’s automated testing found two high-severity jailbreaks; four private red teams found none. # AI https:// spaisee.com/article/gpt-6-2-as tra-was-easier-t…