SAN FRANCISCO, April 20, 2023 — Weights & Biases today unveils new LLMOps tool suite to support Prompt Engineers, unlocking the true potential of Large Language Models (LLMs) for massive business impact . A new suite of tools, W&B Prompts, allows machine learning practitioners to rapidly build state-of-the-art, LLM-based applications that intelligently iterate through a series of prompts in real-time to find inputs that lead to key insights. can do.
With the release of GPT-4 and ChatGPT, there has been a flood of interest and investment in LLM and its potential, with endless applications in areas such as search engines, NLP, healthcare, robotics, code generation, and more. Yet very few companies use LLM in a cohesive way to solve real, concrete problems. LLM use cases are complex, requiring domain expertise, rapid and continuous iteration, and large amounts of computing resources.
Companies such as OpenAI, Stability AI, and Cohere, developing some of the world’s most sophisticated LLMs, rely on Weights & Biases as a core part of their machine learning workflows. Productivity and collaboration for the ML Practitioner and his Prompt Engineer with the latest his LLM-focused enhancements that build upon existing products that support creating and fine-tuning LLMs from scratch will improve.
“Weights & Biases is known for helping the world’s most sophisticated organizations build large-scale language models, and the introduction of W&B Prompts will help companies harness the potential of LLM. We’re breaking new ground,” said VP Phil Gurbacki. Weights & Biases product. “No matter how you use LLM, the Weights & Biases platform tracks your activity and gives you the interactive analytics you need to iteratively develop LLM and LLM-powered applications.”
The W&B Prompts tool suite extends Weights & Biases’ best-in-class scalable experiment tracking and is designed specifically with the needs of Prompt engineers in mind. New features and enhancements include:
- Tools for LLM debugging – Trace Timeline, Trace Table, and Model Architecture enhance rapid engineer tracking of experiments to review past results, identify and debug errors, collect results on model behavior, and share learnings with colleagues make it easier.
- W&B Launch to OpenAI Evals – A new integration framework for W&B Launch evaluation jobs and an easy one-click way to evaluate and score prompts through the OpenAI Evals framework directly from the W&B UI.
- W&B table enhancements – UX improvements within W&B Tables designed to better support the analysis and visualization of Prompt Engineers working in LLM.
- W&B Javascript – New integration allows users to record LLM prompts directly from the web application into Weights & Biases using native Javascript code.
W&B Prompts is now available to all Weights & Biases users. For more information, please visit our website.
About weights and biases
Weights & Biases is a leading developer-first MLOps platform that provides enterprise-grade, end-to-end MLOps workflows to accelerate ML activities. Used by over 500,000 ML practitioners, including teams from OpenAI, Lyft, Blue River Technology, and Toyota, Weights & Biases is part of the new standard for machine learning best practices.
Source: Weights and Bias
