正在加载视频...

视频加载失败

Introducing DeepSeek-V4.1-Flash for understanding research papers 🚀 Highlight any section of a paper to ask questions and “@” other papers for quick context, comparisons, and benchmark references Check it out on any paper on

11,782 次观看 • 2 天前 •via X (Twitter)

3 条评论

Kisson 的头像
Kisson2 天前

When two papers report the same benchmark number under different eval protocols, does the @ comparison flag the methodology mismatch, or surface both numbers side by side as if they're comparable?

Marius Laurusevicius 的头像
Marius Laurusevicius2 天前

Cross paper benchmark comparison is the hard part: the same benchmark name often hides different settings, pass@1 versus best of n, different harness. Does the answer surface the eval config from each paper, or just the headline number?

Rolezn 的头像
Rolezn2 天前

Using DeepSeek-V4.1-Flash to explain DeepSeek-V4.1-Flash 🤯

相关视频