Blog
Read Now
Our in-house series of large action models.
At Salesforce AI Research, we're pioneering the development of xLAM, the next generation of large action models. Our approach involves collecting and synthesizing extensive function/API calling data from open-domain environments and our in-house simulators. This data is then cleaned, standardized, and enhanced to ensure top quality for model tuning.
What is a Large Action Model (LAM)?
Large Action Models are advanced large language models designed to enhance decision-making and translate user intentions into executable actions that interact with the world. They have the potential to automate workflow processes across various domains, making them invaluable for a wide range of applications.
Meet Our Family of xLAM Models
Tiny
xLAM-1B
A compact "Tiny Giant” version of xLAM featuring 1B parameters, perfect for mobile applications.
Small
xLAM-7B
A 7B model designed for swift academic exploration with limited GPU resources.
Medium
xLAM-8x7B
An 8x7B mixture-of-experts model, ideal for industrial applications striving for a balanced combination of latency, resource consumption, and performance.
Large
xLAM-8x22B
A large mixture-of-experts model if you have great computational resources and want to pursue the best performance.
Resources to learn more about xLAM
Blog
Read NowFrom Words to Actions: Introducing xLAM, Salesforce AI Research’s family of Large Action Models
Blog
Read NowThe Rise of Large Action Models Heralds the Next Wave of Autonomous AI
Link
Try it outView our xLAM models on HuggingFace
Paper
Read NowxLAM: A Family of Large Action Models to Empower AI Agent Systems
