Percepta

Percepta is an AI research company focused on advancing LLM capabilities and transformer inference optimization. Their research includes work on executing programs inside transformer models using logarithmic-scaling attention, pushing the boundaries of what large language models can compute efficiently.