As coding agents become more capable, developers increasingly use them to fix bugs through multi-turn interactions. Existing benchmarks do not adequately cov…
机构:腾讯
来源:arXiv 2610.29822 | AI4Papers 论文推荐平台