SantaCoder: don't reach for the stars!
arXiv:2301.03988
Abstract
The BigCode project is an open-scientific collaboration working on the responsible development of large language models for code. This tech report describes the progress of the collaboration until December 2022, outlining the current state of the Personally Identifiable Information (PII) redaction pipeline, the experiments conducted to de-risk the model architecture, and the experiments investigating better preprocessing methods for the training data. We train 1.1B parameter models on the Java, JavaScript, and Python subsets of The Stack and evaluate them on the MultiPL-E text-to-code benchmark. We find that more aggressive filtering of near-duplicates can further boost performance and, surprisingly, that selecting files from repositories with 5+ GitHub stars deteriorates performance significantly. Our best model outperforms previous open-source multilingual code generation models (InCoder-6.7B and CodeGen-Multi-2.7B) in both left-to-right generation and infilling on the Java, JavaScript, and Python portions of MultiPL-E, despite being a substantially smaller model. All models are released under an OpenRAIL license at https://hf.co/bigcode.
Cited by in corpus (8)
- Large Language Models for Code: Security Hardening and Adversarial Testing
- Natural Language Generation and Understanding of Big Code for AI-Assisted Programming: A Review
- LLM for SoC Security: A Paradigm Shift
- Kartezio: Evolutionary Design of Explainable Pipelines for Biomedical Image Analysis
- Metamorphic Malware Evolution: The Potential and Peril of Large Language Models
- Cracks in The Stack: Hidden Vulnerabilities and Licensing Risks in LLM Pre-Training Datasets
- Multi-granular Software Annotation using File-level Weak Labelling
- Deduplicating and Ranking Solution Programs for Suggesting Reference Solutions