Anti GPT - definition of generated texts
{{ info }}
Uniqueness {{ antiGPT['uniqueText'] }}%
Full take. Such a text has already been written! Suspicion of using a neural network! Unsatisfactory. The text needs revision. Suspicion of using a neural network! Satisfactorily. It is desirable to modify the text. Suspicion of using a neural network! Okay. Suitable for most texts. Excellent. The text is unique.
Characters total: {{ сharactersFull }}
{{ sentence['sentence'] }}{{ sentence['sentence'] }}
- {{ error }}
{{ infoService }}
Chrome plugin
Plagiarism checker
Чат с нейросетью
Переводчик
Перефраз
Synonymizer (selection of synonyms)
Title text
Place commas
Spellchecking
Plagiarism checker
Parse
Phonetic word parsing
Perform word parsing by composition
Match words of the same root
Stress the word
Finding the main idea
Anti GPT service helps to identify texts generated by neural network models, rewritten texts without novelty of thought.
How does the service work?
- Neural network models generate texts based on the texts on which they are trained — information for training is taken from open sources and, as a rule, is outdated. The service defines texts generated from open sources.
- Anti-GPT uses a specially trained model to determine the novelty of thoughts in a text.
- Anti-GPT also attempts to identify texts that are machine translated from another language, as the translation does not increase the uniqueness of the meaning of the text.
Distinguishing features of generated texts:
- Uniform uniqueness of meaning across texts, as a rule, no higher than 50%
- Some very non-unique offers
The generated test will be mostly in gray color, the rule gets "Satisfactory. It is desirable to finalize the text.".
Results are stored in history 7 days