"medium" Lms Of Code In The Era Of Llms: Lessons From Stackoverflow · The Large Language Model Bible Contribute to LLM-Bible

The Large Language Model Bible

Research on Large Language Models. Maintained by Sean Moran.

Contact Sean Moran about this survey or website. Made with Jekyll and Hyde.

"medium" Lms Of Code In The Era Of Llms: Lessons From Stackoverflow

Mukherjee Manisha, Hellendoorn Vincent J.. Arxiv 2023

[Paper]    
Applications BERT GPT Model Architecture Multimodal Models RAG Reinforcement Learning Training Techniques

Large pre-trained neural language models have brought immense progress to both NLP and software engineering. Models in OpenAI’s GPT series now dwarf Google’s BERT and Meta’s RoBERTa, which previously set new benchmarks on a wide range of NLP applications. These models are trained on massive corpora of heterogeneous data from web crawls, which enables them to learn general language patterns and semantic relationships. However, the largest models are both expensive to train and deploy and are often closed-source, so we lack access to their data and design decisions. We argue that this trend towards large, general-purpose models should be complemented with single-purpose, more modestly sized pre-trained models. In this work, we take StackOverflow (SO) as a domain example in which large volumes of rich aligned code and text data is available. We adopt standard practices for pre-training large language models, including using a very large context size (2,048 tokens), batch size (0.5M tokens) and training set (27B tokens), coupled with a powerful toolkit (Megatron-LM), to train two models: SOBertBase, with 109M parameters, and SOBertLarge with 762M parameters, at a budget of just \(\)187\( and \)\(800\) each. We compare the performance of our models with both the previous SOTA model trained on SO data exclusively as well general-purpose BERT models and OpenAI’s ChatGPT on four SO-specific downstream tasks - question quality prediction, closed question prediction, named entity recognition and obsoletion prediction (a new task we introduce). Not only do our models consistently outperform all baselines, the smaller model is often sufficient for strong results. Both models are released to the public. These results demonstrate that pre-training both extensively and properly on in-domain data can yield a powerful and affordable alternative to leveraging closed-source general-purpose models.

Similar Work