In Line with Context: Repository-Level Code Generation via Context Inlining
Chao Hu, Wenhao Zeng, Yuling Shi, Beijun Shen, Xiaodong Gu · 2026
Read In Line with Context: Repository-Level Code Generation via Context Inlining by Chao Hu, Wenhao Zeng, Yuling Shi, Beijun Shen, Xiaodong Gu (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-00376.txt.
Description
Chao Hu, Wenhao Zeng, Yuling Shi, Beijun Shen, and Xiaodong Gu tackle code generation for whole software repositories, not just single functions. A model writing code inside a large project must grasp links across functions, classes, and modules. Their 2026 paper notes that retrieval methods lean on surface similarity and often miss those deep links. Their system, InlineCoder, takes a different route.
First the system drafts a rough completion, called an anchor, for the unfinished function. That draft lets it judge confidence and drive a two-way process. The anchor is placed into its callers to show usage, while its own callees are pulled in as context. The combined view turns a hard repository task into something closer to function-level coding.
Catalog record
| CREATOR | Chao Hu, Wenhao Zeng, Yuling Shi, Beijun Shen, Xiaodong Gu |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 54 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2601-00376