rest-stress-testing-reasoning

【开发与工程】Evaluate large reasoning models under stress by asking multiple problems simultaneously, revealing true multi-tasking capacity and context management. Use to identify robustness gaps invisible in single-problem evaluations and discriminate model capabilities beyond traditional benchmarks.

独立 GitHub 快照 · 自动分类 · 以原作者文档为准