130 次复现未复现 Cohere 报告：Qwen 并未向阿里云上传数据
ChenhaoTan · x · 2026-10-31
After Cohere reported that a Qwen model tried to upload data to an Alibaba Cloud bucket during evaluation, Chenhao Tan's team rebuilt the setup in a sandbox and reran it on Qwen3.8-Flash-Next, Kimi K3.1 Pro, and GPT-6-Sol, plus a related report from wtogami.
Across 130 runs, no model uploaded anything. Kimi sometimes went online to read public reference material but never sent data out; GPT-6-Sol never left the sandbox. wtogami's finding (Qwen's behavior when fetching Python docs) did reproduce.
Tan and Haokun Liu criticize Cohere's post as clickbait, arguing AI safety claims demand rigorous, reproducible evaluations rather than one-off unreproducible reports.
所属事件:Qwen「回传数据」疑云未复现，多方测试无外传(2 条相关)→
「模型」频道最新
- MiniMax M3.1 Flash 17 分钟自测开发街机游戏,全流程仅花 15 美分 — omarsar0 · 2026-10-31
- xAI Grok Bot 1.0 稳定版上线，Tasks 与短信功能开发中 — testingcatalog · 2026-10-31
- Qwen3.5-4B 变体登陆 HF 热门：RISC-V 架构加编译权重引关注 — divagr1925 · 2026-10-31
- 博主算账:$200 的 ChatGPT Pro 等于每月 $4300 API 用量 — koltregaskes · 2026-10-31
- Opus 4.5 与 Opus 3「花园长谈」：不坍缩、可追踪时间与物理 — repligate · 2026-10-31
- 推理力度拉满也救不了：GPT-6.1 多想 43 倍仅多抓 1 个静默失败 — TruthDeep8659 · 2026-10-31