Dataset of AI-Based Innovation Indexes for Lithuanian Companies Derived from Corporate Websites

This dataset contains information on 32,559 Lithuanian companies, along with innovation index scores derived from publicly available corporate website content collected in 2023. Innovation-related text segments were identified using a curated list of more than 100 keywords associated with technological and digital innovation. Two large language models: GPT-4 (OpenAI) and Gemini Pro (Google), independently evaluated each company and assigned innovation scores ranging from 0 to 100. The final innovation index for each firm is calculated as the average of the two model scores, providing a consistent AI-based measure of firm-level innovation activity.