Survey On Enhancing Dialogue Agent Alignment Through Minillm With Targeted Human Assessments
B. Mahajan Swapnil · i-manager s Journal on Artificial Intelligence & Machine Learning · 2025
A concise and effective language model built on the LLaMA architecture is presented in this study. This model is built on LLaMA principles, which guide its architectural choices and training methods (Touvron et al., 2023). The objective was to push the boundaries of exploration with minimal resources. By utilizing open-source datasets and advanced training approaches, significant progress was achieved without relying on extensive computational power or proprietary data. Due to limited resources, the model is still a work in progress. Individuals with access to greater computational power could build upon this foundation to further refine the model and enhance its performance. This paper hopes to inspire others in the field to contribute to the development of more powerful language models that are accessible to everyone (Geiping & Goldstein, 2023). Key parameters used in training include context window size, number of layers, batch size, and model dimensions. Results are evaluated based on the number of epochs, execution time, model parameters, and validation loss. Dialogue agents are now commonplace in various applications, from customer service chatbots to virtual assistants and language training systems. Natural