GPT-3.5 Turbo gains fine-tuning option for specific tasks
ChatGPT Gets New Features for Business Use
ChatGPT, an AI chatbot creator, claims that with its new features, it will be a more useful tool for businesses. The GPT-3.5 Turbo version now includes the option for fine-tuning and customization for specific tasks and jobs. For example, it can be directed to improve programming code writing or maintain a more consistent tone during conversations.
This means that companies will now be able to adjust the GPT-3.5 Turbo version of the bot using their own business data. This customized chatbot reportedly can match and even surpass the capabilities of their most advanced GPT-4 language model version in certain tasks.
This way, the bot can become unique to a particular company, offering reliable answers in a specific language or more concise formulations. However, this version lacks the timeliness of other information. Before inputting business data, the model will only have access to information available until September 2021.
On the other hand, OpenAI claims that none of the data used for business purposes will be used for further model training outside of the company using it. Its future use can also include providing a specific branding or marketing ad approach.
The GPT-3.5 Turbo Version for Business Users
Shutterstock / kovopThe GPT-3.5 Turbo model is specifically optimized for use beyond casual conversations with artificial intelligence, as explained by the company. It can process up to four thousand tokens at the same time, which is twice as much as previous models could handle. Testers of the new features reportedly managed to ask 90 percent shorter queries and received good answers after fine-tuning was implemented.
OpenAI, although a pioneer in popularizing artificial intelligence, is not the first to offer GPT models that can be used and trained specifically for business services. Microsoft also offers such capabilities within its AI Builder and Power Virtual Agents services, designed to connect with a company's internal data to provide better responses.