2. Home
  3. ARCHIVE
  4. 2023
  5. VOL 45, NO 2 (2023)
  6. pp. 44-60

CHATGPT AS EXPERTLESS TEST GENERATOR

A.O. Taranowski, V.D. Samoylov

Èlektron. model. 2023, 45(2):44-60

https://doi.org/10.15407/emodel.45.02.044

ABSTRACT

Since its launch there has been a plethora of publications about what ChatGPT can and can’t be used for. While OpenAI remained silent about the exact place of ChatGPT in its framework of language models. Just ‘Try ChatGPT’ appeared to be not the best way to consider its applications for less evident purposes that just composing an essay. The author has strong belief that knowing what something really is adds a lot to one’s understanding of what that something is capable of. This article therefore first considers what ChatGPT really is following its missing manual. That allows than overcoming some crucial limitations of ChatGPT in order to propose, substantiate, and experimentally support its applicability to multiple-choice test items construction on the way to increase automation and optimisation in knowledge assessment systems in order to reach the level of that being expertless. The approach proposed is heavily based upon GUI client for ChatGPT while benefits of being used through API are also explained in the light of integration into existing knowledge assessment systems, additional effects, and future transition to GPT-4.

KEYWORDS

artificial intelligence, large language models, knowledge assessment.

REFERENCES

  1. Bulakh, I.Ye., Mruha, M.R. (2006), Stvoriuiemo yakisnyi test [Creating a quality test], Maisterklas, Kyiv, Ukraine [in Ukrainian]
  2. Abramovych,R.P. (2020), Metody ta zasoby konstruiuvannia kompiuternykh system pidgotovky operatyvno-dyspetcherskogo personal nyzhchykh rivniv v energetytsi. Dysertatsiya na zdobuvannia naukovogo stupenia kandidata tekhnichnykh nauk: 05.13.05 [Methods and tools of designing computer aided systems for training lower levels operational and dispatcher personnel in Energy Engineering. Dissertation for the candidate of technical science degree: 05.13.05], Kyiv [in Ukrainian].
  3. “IntroducingChatGPT”, OpenAI, available at: https://openai.com/blog/chatgpt (accessed 8 April 2023).
  4. “What is ChatGPT?”, OpenAI, available at: https://openai.com/en/articles/6783457-what-is-chatgpt (accessed 8 April 2023).
  5. “Models”, OpenAI, available at: https://platform.openai.com/docs/models (accessed 8 April 2023).
  6. “Model index for researchers”, OpenAI, available at:https://platform.openai.com/docs/model-index-for-researchers (accessed 8 April 2023).
  7. “Product”,OpenAI, available at: https://openai.com/product (accessed 8 April 2023).
  8. “What is ChatGPT Plus?”, OpenAI, available at:https://help.openai.com/en/articles/6950777-what-is-chatgpt-plus (accessed 8 April 2023).
  9. “Introducing ChatGPT and Whisper APIs”, OpenAI, available at: https://openai.com/blog/introducing-chatgpt-and-whisper-apis (accessed 8 April 2023).
  10. “What is ChatGPT?”,OpenAI, available at: https://platform.openai.com/docs/chatgpt-education/what-is-chatgpt (accessed 8 April 2023).
  11. “Large language model”, Wikipedia, available at: https://en.wikipedia.org/wiki/Large_language_model (accessed8 April 2023).
  12. Gozalo-Brizuela, and Garrido-Merchán, E.C. (2023), ChatGPT is not all you need. A State of the Art Review of large Generative AI models, available at: https://arxiv.org/pdf/2301. 04655.pdf (accessed9 April 2023).
  13. “GPT‑4”, OpenAI, available at: https://openai.com/research/gpt-4 (accessed 8April2023).
  14. “An important next step on our AI journey”, available at: https://blog.google/technology/ai/bard-google-ai-search-updates (accessed9 April 2023).
  15. “OpenAI and Microsoft extend partnership”, OpenAI, available at: https://openai.com/blog/openai-and-microsoft-extend-partnership (accessed9 April 2023).
  16. “Microsoft and OpenAI extend partnership”, Microsoft, available at: https://blogs.microsoft.com/blog/2023/01/23/microsoftandopenaiextendpartnership (accessed9 April 2023).
  17. “Terms of use”,OpenAI, available at: https://openai.com/policies/terms-of-use (accessed 9 April 2023).
  18. Samoilov,D., Abramovich, R.P., Lepatiev A.O. (2020), “Computer technologies for the development of training systems for the energy industry”, Elektronne modelyuvannya, Vol. 42, no. 3, pp. 89-97.
    https://doi.org/10.15407/emodel.42.03.089
  19. “Elekrtyschnyi generator” [Electric generator], Wikipedia, available at: https://uk.wikipedia.org/wiki/Електричний_генератор (accessed9 April 2023).
  20. “Electric generator”, Wikipedia,available at: https://en.wikipedia.org/wiki/Electric_generator (accessed 9 April 2023).
  21. Illustration to multimedia training course,Computer‑based personnel training courses ASOT, available at: http://asot.com.ua/img/portfolio/mtc/mtc_screenshot_4.png (accessed 10 April 2023).
  22. Multimedia training courses and banks of test questions,Computer‑based personnel training courses ASOT, available at: http://asot.com.ua (accessed 10 April 2023).
  23. Zakon Ukraiyny “Pro vykorystannia Iadernoiy tnergiiy ta radiatsiinu bezpeku” [Law of Ukraine On Use of Nuclear Power Use and Radiation Safety], Article 22, available at: https://zakon.rada.gov.ua/laws/show/39/95-вр [in Ukrainian] (accessed 10 April 2023).

Full text: PDF