LingmaTongyi / Codev-Bench
Codev-Bench (Code Development Benchmark), a fine-grained, real-world, repository-level, and developer-centric evaluation framework. Codev-Bench assesses whether a code completion tool can accurately capture a developer's immediate intent and suggest appropriate code snippets across diverse, fine-grained contexts.
☆38Updated 4 months ago
Alternatives and similar repositories for Codev-Bench:
Users that are interested in Codev-Bench are comparing it to the libraries listed below
- The repository for paper "DebugBench: "Evaluating Debugging Capability of Large Language Models".☆68Updated 8 months ago
- Official implementation of paper How to Understand Whole Repository? New SOTA on SWE-bench Lite (21.3%)☆72Updated 4 months ago
- ☆22Updated 6 months ago
- CrossCodeEval: A Diverse and Multilingual Benchmark for Cross-File Code Completion (NeurIPS 2023)