Experimenting with multiple AI tools
In the previous lab, we initially received some useful responses from a chatbot. The Poe account that you leverage allows for different models to be tested with your bot to see which best fits your needs and your word choice styling.
Some choices that you may have seen include the following:
- GPT-3.5
- GPT-4
- Claude-instant
- Claude2-100k
- Llama 2
As usual with any development iteration, you should try out different models that best fit the needs of your organization using the same test parameters. In Poe, it is trivial to swap the base bot engine. If you decide to move to another platform for testing, ensure that you are also using the same testing constraints, such as initial prompt structures, external references, and temperature values in the responses.
Our usage of AI shouldn’t be limited to only LLMs in the form of chatbots. Another tool to introduce is SOC Prime’s Uncoder AI. The overall platform...
