Composing alt text using large language models: dataset in Russian

Published: 17 June 2024| Version 1 | DOI: 10.17632/73dptbyxbb.1
Contributors:
Yekaterina Kosova,

Description

The dataset contains the results of developing alternative text for images using chatbots based on large language models. The study was carried out in April-June 2024. Microsoft Copilot, Google Gemini, and YandexGPT chatbots were used to generate 108 text descriptions for 12 images. Descriptions were generated by chatbots using keywords specified by a person. The experts then rated the resulting descriptions on a Likert scale (from 1 to 5). The data set is presented in a Microsoft Excel table on the “Data” sheet with the following fields: record number; image number; chatbot; image type (photo, logo); request date; list of keywords; number of keywords; length of keywords; time of compilation of keywords; generated descriptions; required length of descriptions; actual length of descriptions; description generation time; usefulness; reliability; completeness; accuracy; literacy. The “Images” sheet contains links to the original images. Data set is presented in Russian.

Files

Categories

Text Editing, Accessibility Issue, Image Analysis, Neural Network, ChatGPT, Chatbot

Licence