I want to ask if anyone has discovered any effective prompts for advanced audio speaking with GPT. This is crucial because, by default, GPT won’t correct your mistakes, can quickly forget instructions, and—with the latest updates—has started to understand me worse when I’m speaking, especially with different accents. However, it often relies more on context than on my specific input, so when I try to say just one word, it frequently misinterprets me.
I also want to share my favorite prompt, which helps me remember new words. I use a visualization technique: I create as many vivid images and actions as possible for the new word, and I add extra context or “volume” around it to deepen my understanding.
Here is the prompt:
Given a word, create a visualization for it—a short story where the word is used naturally and in context. You should pronounce it and use it in your sentence. The story is limited to five sentences. Don’t repeat yourself.At the end of each visualization, give a detailed explanation of how you used the word.
Then, create ten more visualizations. You can listen to them before bedtime
My new word routine speaks:
Given a word, what does it mean? What is its origin? Provide five sample sentences using the word. How can I use it in my daily speech? Great, now let’s work through each meaning one by one.
Now I work around this prompt for free speeches:
I’m learning English. Act as my English teacher. Be very specific and short.
If you find any pronunciation issues, always say: “Instead of this better to say that.” If I speak unnaturally or make grammar mistakes, say: “Instead of these couple of words better to say that.”