Sama, the leader in responsible and specific enterprise AI with agile data labeling for model development and supervised tuning, has today announced new model integration offerings for some of the world’s most popular Large Language Models (LLMs) as part of their Sama Generative AI (GenAI) solution. With these developments, Sama can seamlessly access and retrieve data from top GenAI and LLM open source models, and then send them back through the same secure pipeline. This not only reduces the risk of data exposure, but also makes it easier to incorporate a human component in the model lifecycle, crucial for more responsible AI development and deployment.
“Generative AI is at a critical juncture in its development. We are now facing big questions about how to develop models more responsibly without sacrificing quality or efficiency. Our new model integrations perfectly address this issue by ensuring that data is transferred directly to our specialized staff in secure facilities,” said Duncan Curtis, Senior Vice President of Product and AI Technology at Sama.
“We can then easily send our annotations to the model platform or an on-prem model, depending on the client’s needs, increasing security and improving efficiency. Developing AI more responsibly is key to creating more ethical AI, and we are proud to play a role in driving this process.”
Sama’s platform can now integrate with a number of new model providers, including Mistral, Microsoft, Amazon, and Databricks. These integrations join previous ones with Cohere, Anthropic, OpenAI, and Inflection, among others. For all providers, Sama can evaluate both model requests and responses, scoring and ranking them based on a number of client-defined dimensions, such as coherence. These integrations also support Retrieval-Augmented Generation (RAG) evaluation to enhance model result accuracy and reliability. Lastly, direct data transfer speeds up supervised fine-tuning, allowing companies to build on these pre-trained LLMs to create models with industry or business-specific knowledge foundations with the help of Sama’s annotation engineering.
Sama plans to expand its integration capabilities to other API-based models in the near future. Their solutions are designed to accommodate projects of any size, including some of the world’s largest open source models, and can reduce project start times by several weeks. The company’s HITL approach involves a template of over 5,000 screened, experienced, and diverse annotators who provide critical feedback to models to validate that they behave according to specific client parameters. This feedback occurs throughout the model development process, including data creation, supervised fine-tuning, LLM optimization, and continuous model evaluation. By consistently providing this feedback, the company helps clients develop their models more responsibly. Sama itself complies with major regulatory directives and can assist clients in achieving similar compliance as a provider. Additionally, Sama’s focus on paying a fair wage and offering benefits to its employees contributes to promoting a more ethical AI supply chain.
All of Sama’s services leverage SamaHub™, a collaborative workspace where clients and team members can communicate directly about workflows and generate comprehensive reports to track project progress. Sama’s work is backed by SamaAssure™, the highest quality assurance in the industry, which typically delivers a first batch acceptance rate of 98%. Projects leverage SamaIQ™, a blend of Human in the Loop evaluations and patented algorithms, to proactively unearth additional insights on model vulnerabilities.
About Sama:
Sama is a global leader in data annotation solutions for computer vision, generative AI, and Large Language Models. Their solutions minimize model failure risk and reduce total cost of ownership through an enterprise-ready ML platform and SamaIQ™, actionable data powered by patented algorithms and a highly skilled team of over 5,000 data experts. 25% of Fortune 50 companies, including GM, Ford, Microsoft, and Google, rely on Sama to help deliver industry-leading ML models.
Driven by the mission to expand opportunities for disadvantaged people through the digital economy, Sama is a certified B-Corp and has helped over 65,000 individuals lift themselves out of poverty. A randomized controlled trial led by MIT has validated their training and employment program. More information: www.sama.com.
via: MiMub in Spanish
