The references found by the chatbots.

This dataset contains the raw data used in the study evaluating the accuracy of bibliographic references generated by large language model–based chatbots. Six chatbot configurations were tested: ChatGPT (OpenAI, USA) - versions 4.0, 4.1, and 4.1 with web browsing enabled, Gemini (Google, USA) - versions 1.5 Pro, 2.5 Pro, and 2.5 Pro with web search. Each chatbot was asked to generate references for a predefined set of prompts related to auditory research. The dataset includes all responses- 1947 references.