Abstract
This study aims to analyze the level of knowledge, attitudes, and perceptions of elementary school teachers toward the implementation of deep learning at Cluster 1 Labuapi and to examine the relationships among these variables. This research employed a quantitative correlational design with 67 teachers selected through proportional random sampling. Data were collected using a validated knowledge test and Likert-scale questionnaires, with reliability confirmed by Cronbach’s Alpha. Data were analyzed using descriptive and inferential statistics. The results show that teachers have a high level of readiness, with average scores of 83.75% (knowledge), 75.68% (attitude), and 76.48% (perception). However, the relationships among these variables are positive but very weak and statistically insignificant, with R² values below 0.05, indicating that high knowledge does not necessarily strengthen attitudes and perceptions. This finding is influenced by the gap between conceptual understanding and practical implementation. Teachers’ attitudes and perceptions are more affected by external factors, such as limited practical experience, instructional constraints, and insufficient institutional support. As a result, knowledge remains at the cognitive level and is not fully reflected in attitudes and perceptions. Therefore, improving deep learning implementation requires not only enhancing teachers’ knowledge but also strengthening practical experience and institutional support.