A Systematic Evaluation of Trajectory Data Curation for LoRA Fine-Tuning of Code Agents
A systematic empirical study of trajectory data filtering for LoRA fine-tuning of Qwen2.5-Coder-7B-Instruct on the SWE-trajectory dataset and a two-axis quality scoring framework are presented, establishing trajectory-level quality scoring as a viable but scale-sensitive lever for code-agent SFT.