Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{194697,
        author = {C. JASMINE SHIRELY and Dr. M.VARGHEESE},
        title = {A FRAMEWORK FOR ADAPTIVE PROMPT OPTIMIZATION AND ROBUSTNESS EVALUATION IN GENERATIVE AI MODELS},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {10},
        pages = {4937-4941},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=194697},
        abstract = {Large Language Models (LLMs) such as GPT-4 and LLaMA-3 have become foundational in the era of generative artificial intelligence, enabling diverse applications in reasoning, dialogue, and knowledge synthesis. However, the quality and robustness of their responses are highly sensitive to the phrasing and structure of input prompts. This dependency raises the challenge of optimizing prompts to ensure accuracy, consistency, and resilience against adversarial or noisy inputs.
This work proposes an Adaptive Prompt Optimization Framework (APOF) that dynamically refines input prompts through a reinforcement-based feedback loop, embedding analysis, and robustness testing. The framework integrates semantic similarity evaluation, reinforcement learning–based optimization, and a robustness scoring mechanism to produce stable, high-quality generations. Experimental evaluation across GPT-4 and LLaMA-3 demonstrates that the proposed system improves contextual accuracy by approximately 18% and robustness to linguistic perturbations by 22% compared to baseline prompting strategies.},
        keywords = {Large Language Models, Prompt Optimization, Generative AI, Robustness, Reinforcement Learning, GPT-4 L LaMA.},
        month = {March},
        }