PulseAugur
实时 17:07:36
English(EN) The root cause: GPQA publishes its eval set in a single split labeled 'train' on Hugging Face, and Soofi's pipeline selected by name rather than meaning. An aud

AI 基准测试污染可追溯至管道错误;分数被撤回

一个管道错误导致包括 GPQA 在内的四个 AI 基准测试受到污染。该问题源于 GPQA 评估集在 Hugging Face 上以 'train' 标签发布,导致一个管道根据名称而非预期含义错误地选择了它。这一疏忽已导致这些基准测试的分数被撤回,未来的数据集将 undergo allowlist validation and n-gram screening 以防止再次发生。 AI

影响 基准测试数据集中的错误会扭曲性能评估,可能误导研究和开发工作。

排序理由 该项目讨论了 AI 基准测试数据集和评估管道的问题,这属于研究范畴。[lever_c_demoted from research: ic=1 ai=1.0]

在 Mastodon — fosstodon.org 阅读 →

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

AI 基准测试污染可追溯至管道错误;分数被撤回

报道来源 [1]

  1. Mastodon — fosstodon.org TIER_1 English(EN) · [email protected] ·

    The root cause: GPQA publishes its eval set in a single split labeled 'train' on Hugging Face, and Soofi's pipeline selected by name rather than meaning. An aud

    The root cause: GPQA publishes its eval set in a single split labeled 'train' on Hugging Face, and Soofi's pipeline selected by name rather than meaning. An audit found the same failure across four benchmarks. Future datasets now get allowlist validation and n-gram screening. htt…