Data and materials for "LLM-Based Semantic Support in SAE Level 3 Automated Driving: Supervisory Responses and Intervention Decisions"

Résumé fourni par la source

This dataset accompanies a video-based study of LLM-based semantic support during automated-driving supervision. It contains de-identified behavioral and questionnaire data from 32 participants, Chinese experimental messages with English translations, message characteristics and audio-duration summaries, model specifications and saved results, Supplementary Tables S1–S6, and Figures 4–8. The package includes Python code for recalculating the reported RQ3 signal-detection contrasts from the saved model fit. It does not include the complete model-fitting software or the original DrivingDojo videos. The README and data dictionary describe the files, variables, and analysis scope.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.