Discover the Best LLM for your Automated Labeling: Automate NLP Labeling with Multiple LLMs Simultaneously

We’re thrilled to unveil our latest feature: LLM Assisted Labeling! With this feature you will be able to simultaneously compare multiple LLMs and their ability to automatically label your NLP projects. This powerful new update integrates various Large Language Model (LLM) providers into our ML Assisted Labeling extensions, making your work more efficient and accurate.

What’s New?

Previously, our ML-Assisted Labeling could process only one input column at a time. Now, with LLM-Assisted Labeling, you can use multiple columns, allowing for more accurate model predictions and broader use cases.

Evolution of ML-Assisted Labeling

Datasaur’s ML Assisted Labeling feature has already streamlined the labeling process by leveraging various providers, including early access to LLM-based tools such as OpenAI and LLM Labs. Previously, the ML-assisted OpenAI provider supported direct API calls to OpenAI GPT models, allowing users to directly adjust prompts and settings (like temperature and max tokens) in the project interface. On the other hand, the ML-assisted LLM Labs provider facilitated API calls from any of the 200+ LLMs within Datasaur’s LLM Labs, supporting various LLM providers though not offering in-project adjustment capabilities. Access to both OpenAI and LLM Labs providers was limited to a closed beta.

Introducing LLM-Assisted Labeling

Based on invaluable feedback from our users, we are excited to introduce the new and improved LLM Assisted Labeling, combining the strengths of both OpenAI and LLM Labs providers.

LLM Assisted Labeling offers users a broader range of LLM options and unparalleled flexibility to experiment and determine the best LLM provider and setup for their data labeling needs.

Start in 3 Simple Steps

• ‍Set up your labeling project: Begin by creating a row or span-based labeling project in Datasaur. This setup is straightforward and user-friendly.‍
• Configure LLM Assisted Labeling settings: Select the ML Assisted extension, choose LLM Assisted Labeling, and configure the settings and prompt

‍

• Predict and Review: Select specific rows or all rows. Click “Predict label” to view LLM-assisted label recommendations. Experiment with selected data, prompts, LLM providers, and settings to find the ideal labeling assistant for your specific data.

For a detailed guide, please visit our documentation.

Benefits of Using LLM-Assisted Labeling

• ‍Access to Top LLMs: With LLM-Assisted Labeling, you can test various top LLMs with different settings and prompts for labeling—all within one convenient page.
• Samless Integration: Adjust prompts and settings like temperature and max tokens in real-time within the Datasaur workspace. Test configurations on selected rows without affecting the entire dataset.

Transform Your Labeling Workflow Today

At Datasaur, we are committed to making your labeling process as efficient and accurate as possible. Our LLM-Assisted Labeling feature allows you to access, compare, and iterate on multiple LLMs, transforming traditionally time-consuming tasks into streamlined, enjoyable experiences.

Ready to elevate your data labeling? Get started with Datasaur today! Contact us at support@datasaur.ai for more information. Your path to efficient data labeling starts now!