English(EN) Two Accounts, One Vault, and Where the Tokens Actually Went

Anthropic的Claude模型出现意外代币泄露

作者 PulseAugur 编辑部 · [1 个来源] · 2026-09-04 15:41

一项技术分析显示，Anthropic的Claude模型存在显著的代币泄露，这与最初对其位置的假设相反。研究人员发现泄露并非发生在预期的区域，表明模型架构内部存在更复杂的问题。这一发现表明需要进一步调查Claude的内部工作机制，以解决意外的代币支出问题。 AI

影响识别出Claude代币管理中潜在的低效率或漏洞，可能影响运营成本或性能。

排序理由对现有AI模型中特定技术问题的分析。[lever_c_demoted from research: ic=1 ai=1.0]

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

本文如何被排名

Signal score: 33 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 对现有AI模型中特定技术问题的分析。[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: 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。

Medium — Claude tag TIER_1 English(EN) · Dz Anton · 2026-09-04 15:41

两个账户，一个金库，以及代币的实际去向

<div class="medium-feed-item"><p class="medium-feed-snippet">The comments said we had a leak. We went and measured it. They were right, and the leak was not where anyone guessed.</p><p class="medium-feed-link"><a href="https://medium.com/@dzyatkovskiy.a2/two-accounts-one-vault-an…