Google Allows Users To Fine-Tune Gemini For More Precise Responses

Search engine giant Google is rolling out changes to its AI chatbot Gemini so that it can produce more precise replies.

The company announced yesterday how it’s making way for greater means of customization for fine-tuning of the tool. This can now be done while selecting a certain part of the reply to the prompt by pressing on the pencil icon.

This would give rise to a new tab called modify selected text and it would open up new boxes with tabs like Regenerate, Short, Long, And Remove as options to choose from. You can even open the whole text field if you want.

For instance, if your initial prompt was linked to Plan a 3-day trip to New York City that centered around the areas of Central Park, you can see a reply that entails the tool telling you how that’s not possible if you made changes to the draft to include four days instead of 3.

You might get replies like how the tool says that’s not possible a prompt that violated the whole Prohibited Use Policy or something about the tool not comprehending what’s mentioned.

In cases where you feel the prompt is not a part of any such variant, you can always make adjustments to the selected words or even go about rewording the instructions. But do remember that you can’t modify it completely.

What you can do is fine-tune to include any text as a reply from an extension or any text in reply to where the code was utilized for producing results like chart text. And also it would entail any text as a part of the reply that has code blocks. Last but not least, it includes a selection that features all sorts of pictures.

The goal of this endeavor is to assist users in getting results that are similar to what they might be in search of. So what we can see from all of this is that Google wishes to give users greater control over their entire creative process by allowing them to detail content and thoughts in regards to the real reply.

This joins the ability to witness other kinds of drafts or reproduce context that’s short, casual, professional, or longer in length - totally up to you. For now, it’s launching in the English language for those using the web tool.

Read next: Generative AI Continues To Promote Fake Information About US Elections, New Research Proves
Previous Post Next Post