From Completion to Editing: Unlocking Context-Aware Code Infilling via Search-and-Replace Instruction Tuning
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866914265273729024 |
|---|---|
| author | Zhang, Jiajun Cui, Zeyu Yang, Jiaxi Zhang, Lei Jing, Yuheng Ma, Zeyao Bai, Tianyi Wang, Zilei Liu, Qiang Wang, Liang Hui, Binyuan Lin, Junyang |
| author_facet | Zhang, Jiajun Cui, Zeyu Yang, Jiaxi Zhang, Lei Jing, Yuheng Ma, Zeyao Bai, Tianyi Wang, Zilei Liu, Qiang Wang, Liang Hui, Binyuan Lin, Junyang |
| contents | The dominant Fill-in-the-Middle (FIM) paradigm for code completion is constrained by its rigid inability to correct contextual errors and reliance on unaligned, insecure Base models. While Chat LLMs offer safety and Agentic workflows provide flexibility, they suffer from performance degradation and prohibitive latency, respectively. To resolve this dilemma, we propose Search-and-Replace Infilling (SRI), a framework that internalizes the agentic verification-and-editing mechanism into a unified, single-pass inference process. By structurally grounding edits via an explicit search phase, SRI harmonizes completion tasks with the instruction-following priors of Chat LLMs, extending the paradigm from static infilling to dynamic context-aware editing. We synthesize a high-quality dataset, SRI-200K, and fine-tune the SRI-Coder series. Extensive evaluations demonstrate that with minimal data (20k samples), SRI-Coder enables Chat models to surpass the completion performance of their Base counterparts. Crucially, unlike FIM-style tuning, SRI preserves general coding competencies and maintains inference latency comparable to standard FIM. We empower the entire Qwen3-Coder series with SRI, encouraging the developer community to leverage this framework for advanced auto-completion and assisted development. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_13384 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | From Completion to Editing: Unlocking Context-Aware Code Infilling via Search-and-Replace Instruction Tuning Zhang, Jiajun Cui, Zeyu Yang, Jiaxi Zhang, Lei Jing, Yuheng Ma, Zeyao Bai, Tianyi Wang, Zilei Liu, Qiang Wang, Liang Hui, Binyuan Lin, Junyang Software Engineering Computation and Language The dominant Fill-in-the-Middle (FIM) paradigm for code completion is constrained by its rigid inability to correct contextual errors and reliance on unaligned, insecure Base models. While Chat LLMs offer safety and Agentic workflows provide flexibility, they suffer from performance degradation and prohibitive latency, respectively. To resolve this dilemma, we propose Search-and-Replace Infilling (SRI), a framework that internalizes the agentic verification-and-editing mechanism into a unified, single-pass inference process. By structurally grounding edits via an explicit search phase, SRI harmonizes completion tasks with the instruction-following priors of Chat LLMs, extending the paradigm from static infilling to dynamic context-aware editing. We synthesize a high-quality dataset, SRI-200K, and fine-tune the SRI-Coder series. Extensive evaluations demonstrate that with minimal data (20k samples), SRI-Coder enables Chat models to surpass the completion performance of their Base counterparts. Crucially, unlike FIM-style tuning, SRI preserves general coding competencies and maintains inference latency comparable to standard FIM. We empower the entire Qwen3-Coder series with SRI, encouraging the developer community to leverage this framework for advanced auto-completion and assisted development. |
| title | From Completion to Editing: Unlocking Context-Aware Code Infilling via Search-and-Replace Instruction Tuning |
| topic | Software Engineering Computation and Language |
| url | https://arxiv.org/abs/2601.13384 |