The year was 2013. In a state-of-the-art kitchen laboratory in Stanford University’s Robotics Center, surrounded by the whir of servo motors and the aroma of brewing coffee, I observed our latest prototype attempt…
Developers face unique challenges when retrieving code snippets, such as understanding syntax, control flow, and variable dependencies. Enter SFR-Embedding-Code, a groundbreaking family of code embedding models that aims to address these challenges and revolutionize how we retrieve and generate code.
We present TACO, a family of multi-modal large action models designed to improve performance on complex questions that require multiple capabilities and demand multi-step solutions.
To address the challenges in generating multimodal instruction data, we developed ProVision, a scalable, programmatic framework that employs scene graphs and human-written programs to systematically synthesize vision-centric instruction data.
The recent launch of Agentforce marks a pivotal moment in orienting Salesforce and our customers’ businesses toward an AI-empowered future. In this emerging landscape, augmented by a network of AI agents, the role…
From new open source models to evaluation frameworks, our AI Research team has been moving the needle in AI. Take a look at some of our 2024 highlights.
Salesforce's trusted AI architecture for red teaming leverages automation to scale ethical AI testing, utilizing a tool called fuzzai to simulate diverse adversarial scenarios and enhance model robustness. By automating adversarial prompt generation and response validation, fuzzai helps secure AI interactions while reducing human exposure to harmful content.
Now generally available, Agentforce for Developers represents a significant step in Salesforce's mission to drive innovation and deliver intelligent development tools. Let’s explore how Agentforce, powered by Salesforce AI Research’s large language models, is transforming the way you code.
Time series forecasting is becoming increasingly important across various domains, thus having high-quality, diverse benchmarks are crucial for fair evaluation across model families.