Training Code Generation Models for Self-Debugging
from Neural intel Pod Hosted by Neuralintel.org
Amazon Science is focused on improving code generation through debugging. They use large language models (LLMs) to both generate and debug code, leveraging techniques like supervised fine-tuning and reinforcement learning to train the models. A key element involves creating synthetic debugging data to overcome the scarcity of real-world examples. This approach shows significant improvement in code performance as measured by standard benchmarks. The team utilizes methods like chain-of-thought reasoning and focuses on refining models. A variety of job opportunities are also presented in areas including generative AI, transportation optimization, and advertising technology. These are all related to machine learning, data science, and AI within Amazon and AWS.
Embed this episode
No transcript yet.