A Systematic Evaluation of Trajectory Data Curation for LoRA Fine-Tuning of Code Agents
代码智能体LoRA微调的轨迹数据管理系统评估
专题命中 指令微调 :SFT(abstract);分类 cs.AI
AI总结 研究代码智能体LoRA微调中轨迹数据管理,提出效率和风格两轴质量评分框架,经16个控制实验评估,揭示规模依赖的质量-数量权衡及错误重试率是主要子维度,为代码智能体监督微调提供可行手段与评估协议。
Comments 12 pages, 2 figures. Author's accepted manuscript of a paper published in the ICIC 2026 proceedings (oral presentation)
Journal ref In: Huang, D.S., Li, B., Zhang, Q., Bao, W. (eds) Advanced Intelligent Computing Technology and Applications. ICIC 2026. Lecture Notes in Computer Science, vol 16669, pp. 39-50. Springer, Singapore (2027)