arXiv:2602. 06470v3 Announce Type: replace-cross Abstract: Scaling training data and model parameters has long driven progress in large language models (LLMs), but this paradigm is increasingly constrained by the scarcity of high-quality data and diminishing returns from rising computational costs.
Code Comprehensor is a hybrid retrieval‑reasoning framework designed to aid understanding of large codebases. It combines a code dependency callgraph, CodeScribe—which generates hierarchical documentation using code‑aware LLMs—and CodeContrast, a CLIP‑style dual‑encoder that learns aligned text‑code embeddings for fast, semantically precise retrieval. Experiments on OpenCV and a 5M‑LOC proprietary repository show that CodeScribe‑derived supervision improves cross‑modal retrieval accuracy, and LLM‑based evaluation reports over 90% faithfulness.
arXiv:2508. 16131v3 Announce Type: replace-cross Abstract: Code completion entails the task of providing missing tokens given a surrounding context.
arXiv:2607. 24882v1 Announce Type: cross Abstract: Modern coding agents are usually evaluated by whether they eventually produce a correct patch, but patch generation depends on an earlier context-acquisition stage: finding the repository files needed for the task.
arXiv:2609.01601v1 Announce Type: cross Abstract: The repository-level code generation task requires synthesizing code that satisfies task requirements while remaining consistent with the target repo...
arXiv:2608.27831v1 Announce Type: cross Abstract: Coding agents are now commonly evaluated on the SWE-bench family of benchmarks, whose tasks are built from curated GitHub issues--long, structured, a...
arXiv:2606. 03657v1 Announce Type: new Abstract: Large language models for code generation often need to use APIs that are absent from their pretraining data.
arXiv:2610.05437v2 Announce Type: replace Abstract: Computer-use agents need to capture procedural knowledge of how people use software. User telemetry offers a scalable source of this knowledge. How...
arXiv:2606. 14654v1 Announce Type: new Abstract: Sequential or time-stamped interaction logs provide objective records of digital application usage, yet their granularity and noise often obscure meaningful insights into people's work.
The paper introduces SaraCoder, a retrieval‑augmented framework that improves knowledge density for repository‑level code completion by pruning redundant context and selectively injecting cross‑file dependency definitions within a fixed token budget. It employs a Hierarchical Feature Optimization module to reduce an over‑complete candidate pool in four stages, freeing tokens for an External‑Aware Identifier Disambiguator. Experiments on CrossCodeEval and RepoEval‑Updated demonstrate that SaraCoder outperforms strong RAG baselines such as GraphCoder and Repocoder across multiple code LLMs without increasing token usage, and achieves synergistic gains when combined with other cross‑file methods.
arXiv:2608. 10319v1 Announce Type: cross Abstract: Large language model (LLM)-powered agents have rapidly evolved from code-completion tools into solvers of complex software engineering tasks.
The paper introduces DECODE, a dataset of 53.6K real‑world in‑IDE edits of AI‑generated code from over 1,000 developers, and uses it to study when, why, and how developers modify such code. Findings show developers spend little effort editing and often abandon AI outputs, highlighting the need for LLMs to infer intent from edits. Experiments reveal that while frontier models can generate edits that align with developer intent, they tend to over‑edit and revert changes; fine‑tuning improves performance but does not fully close the gap.