Training and Deploying Standalone LLM + Guardrails

Explore the process of training and deploying a standalone Large Language Model (LLM) with guardrails in this engaging demo.

Witness the stark difference between an unguarded LLM generating whimsical, off-brand responses and one with carefully implemented guardrails, maintaining professionalism and coherence across various prompts.

Explore real-world examples, such as crafting LinkedIn posts, where the guardrailed LLM effortlessly captures the appropriate voice and context, while the unguarded model deviates, potentially compromising brand integrity.

Ready to dive in? Gain access to LLM DataStudio and H2O LLM Studio for your training endeavors:

1. Visit our Aquarium platform at https://aquarium.h2o.ai/.

2. Follow the steps in the video tutorial to create an account on Aquarium: Accessing h2o.ai Aquarium Labs.

3. Once you're in Aquarium, navigate to the LLM Data Studio Lab or H2O LLM Data Studio Lab, depending on your requirements.

4. Launch an instance to access the user interface via the provided URL link at the bottom of the page. The instance will remain available for 120 minutes, after which all data will be erased. Enjoy your training session!