DeepSeekCoder
Emerging4papers using it
2024first seen
The 'DeepSeekCoder' is a benchmark dataset used to evaluate the qualitative aspects of automatically generated code, specifically focusing on reliability, maintainability, and security, in the context of model quantization effects.
Papers using DeepSeekCoder (4)
- Understanding Robustness of Model Editing in Code LLMsBridging Developer Instructions and Code Completion Through Instruction-Aware Fill-in-the-Middle ParadigmIs Quantization a Deal-breaker? Empirical Insights from Large Code ModelsMCGMark: An Encodable and Robust Online Watermark for Tracing
LLM-Generated Malicious Code