Rewrite Sentences Examples
In this post, we’re going to produce some rewrite sentences examples using a new feature that is coming to wordbot.io in the near future. The feature will allow you to freeze, or skip, certain words when rewriting an article.
[yop_poll id=”2″]
Skip Rewriting Words for SEO Optimization
The thought is when you’re writing a blog post or article for search engine optimization (SEO), you’re going to be targeting certain keywords or key phrases. As you edit and rewrite your content, the last thing you’ll want is wordbot to rewrite your content and remove your keywords.
The new word-freezing feature will allow you to quickly add (from within the text editor) a list of words to freeze, or skip when rewriting your articles. This will allow you to continue to produce robust rewrites with wordbot while keeping your article fully optimized for SEO for the keywords and key phrases you’re targeting.
Adding a New Feature – Why Rewrite Sentences Examples?
The first step to adding a feature like this is to change our AI. We need to change and/or create new GPT-3 prompts that will allow us to skip words when rewriting sentences. This is no easy feat and one we’re still working on. That’s why we’re going to produce some rewrite sentence examples in this post. These examples allow us to test our AI progress before we make the effort to build the feature into the software.
If we can get the AI to work, we ‘ll then add the feature into the software (wordbot). We’ll probably add the button to the rich text editor toolbar as shown below.
Examples With Word Freezing
For each of the example sentence rewrites, we’re going to list the word(s) we want to freeze, the original sentence, the rewritten sentence without the word freezing feature, and the rewritten sentence with the word freezing feature. We’ll do a handful of sentence examples and then discuss the results at the end of the post. As I write this I don’t know what the results will be, so let’s hope the new AI does a good job or its back to the drawing board.
Rewrite Sentences Examples – 1
Word to freeze: smoke
Original Sentence: He was in the habit of smoking. He no longer smokes.
Rewrites Without Freezing: 100%
Paraphrase 1: He used to smoke. But he doesn’t anymore.
Paraphrase 2: He always smoked, but now he’s not doing it anymore.
Paraphrase 3: He once smoked, but he doesn’t do it anymore.
Rewrites With Freezing: 100%
Paraphrase 1: He was in the habit of smoking. He does not smoke now.
Paraphrase 2: He used to smoke, but now he does not smoke.
Paraphrase 3: He was in the habit of smoking, but now he does not smoke.
Rewrite Sentences Examples – 2
Word to freeze: company
Original Sentence: She joined our company in 2005. She quit our company in 2011.
Rewrites Without Freezing: 100%
Paraphrase 1: She joined our company in 2005, and then she quit our company in 2011.
Paraphrase 2: She has worked for our company for over a decade. (factually incorrect rewrite)
Rewrites With Freezing: 100%
Paraphrase 1: She worked at our company in 2005. Then she left our company in 2011.
Paraphrase 2: She had joined our company in 2005. Then she left our company in 2011.
Paraphrase 3: She has been with our company since 2005. Then she quit our company in 2011.
Rewrite Sentences Examples – 3
Word to freeze: article rewriter
Original Sentence: An article rewriter is a piece of software used to paraphrase content.
Rewrites Without Freezing: 66%
Paraphrase 1: A piece of software called content rewriter is used to re-word content.
Paraphrase 2: A software called article rewriter is used to paraphrase content.
Paraphrase 3: The article rewriter is a piece of software that is used to re-word content.
Rewrites With Freezing: 100%
Paraphrase 1: There is a piece of software called an article rewriter.
Paraphrase 2: An article rewriter is a piece of software that can be used to rewrite content.
Paraphrase 3: There is a piece of software called an article rewriter that can be used in order to rewrite content.
Rewrite Sentences Examples – 4
Word to freeze: lunch, restaurant
Original Sentence: The woman ate lunch at the restaurant
Rewrites Without Freezing: 66%
Paraphrase 1: The chick had had lunch at the place.
Paraphrase 2: The woman had had lunch at the restaurant.
Paraphrase 3: A lady had eaten lunch at the restaurant.
Rewrites With Freezing: 66%
Paraphrase 1: The woman had taken lunch at a restaurant.
Paraphrase 2: The woman had eaten lunch at a restaurant.
Paraphrase 3: The woman had eaten a meal at the restaurant.
Rewrite Sentences Examples – 5
Word to freeze: shopping
Original Sentence: The woman went shopping with her friends.
Rewrites Without Freezing: 66%
Paraphrase 1: The woman went to the mall with her friends.
Paraphrase 2: The woman had gone shopping with her friends.
Paraphrase 3: The woman was going shopping with her friends.
Rewrites With Freezing: 100%
Paraphrase 1: She went on a shopping spree with her friends.
Paraphrase 2: She had gone shopping with her friends.
Paraphrase 3: The woman had gone shopping with her friends.
Conclusion
I wanted to start off small with simplistic, fabricated sentences. The sentences tested are very basic and great for a first round of testing our new gpt-3 prompts to freeze words. As we improve the ai to lock words, we’ll advance to more sophisticated and complex sentences. Let’s discuss the results.
First, let’s point out an inaccuracy. One of Example 2’s rewrites was factually incorrect. It rewrote the sentence to say the lady worked at the company for over a decade, which is not true based on the original sentences. This is a great example of why many experts are concerned about AI – if not fact-checked and used correctly, it can produce swaths of incorrect and misleading information. But I digress, back to the actual sentence rewriting testing.
The AI without the word-freezing mechanism included the frozen SEO keyword in the rewritten sentences 12 times of out 15, or 80% of the time. The AI with the word-freezing mechanism included the frozen SEO keyword in the rewritten sentences 14 times out of 15, or 93%.
A few take aways.
First, our existing technology without the ability to freeze keywords and phrases is still very good at including them in the rewritten content. 80% is not bad at all. However, 93% is much better. This percentage warrants us doing further testing and tweaking on the gpt-3 prompt.
Second, as we add the ability to lock in words we also add the possibility of creating poorer-quality rewrites. The user will have to carefully weight this into their decision when deciding to lock keywords or not.
Example 5 above is a great demonstration of this. We wanted to keep the word “shopping” in the rewrites, but one of the rewrites produced was “The woman went to the mall with her friends”. The rewrite failed in the sense of including the keyword “shopping”, but succeeded in the sense of producing an excellent rewrite of the sentence.
