GENERator is a groundbreaking long-context generative foundation model that leverages large language model capabilities to interpret complex genomic sequences beyond the limitations of existing models.

• Processes long DNA contexts (up to 100k tokens) to capture complex biological relationships
• Achieves state-of-the-art performance on genomic benchmark tasks including protein-coding region prediction
• Demonstrates ability to interpret genetic variations associated with diseases and phenotypes
• Employs an innovative hierarchical tokenization approach to handle genomic data efficiently

This research represents a significant advancement in computational genomics, potentially accelerating medical discoveries by improving our ability to analyze and interpret the human genome at scale.

GENERator: A Long-Context Generative Genomic Foundation Model