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{204304,
        author = {Arunadevi Thirumalraj and S.venkatasubramanian and Subasri.A},
        title = {Bias Reduction and Hallucination Mitigation Techniques in Large-Scale Generative Language Models},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {13},
        number = {1},
        pages = {1621-1629},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=204304},
        abstract = {Modern artificial intelligence applications rely heavily on large-scale generative language models (LLMs), however these models can still suffer from representational distortions, hallucinations, and systemic biases, which make them unreliable, unsafe, and untrustworthy. In this paper, a conceptual-theoretical framework combining mechanisms of bias reduction, hallucination detection architecture, and safe-generation protocol is suggested to fit modern LLMs. Based on the progress in the parameter of model alignment, dataset curation, uncertainty quantification, retrieval-augmented generation, reinforcement learning based on human feedback, and constraint-based decoding the paper posits hallucination mitigation as a design priority that is deeply embedded into the training pipeline. It builds on the idea that bias and hallucination are scale-dependent emergent properties of model design, dynamical strategies of statistical learning, and distributional discrepancies, so to avert them, multi-level interventions such as pre-training data control to post-processing verification cycles will be necessary. Suggestions are made on conceptual indicators that can be used to assess model integrity, factual consistency, contextual grounding and fairness strength.},
        keywords = {ESG Investing; Sustainable Finance; Financial Inclusion; FinTech; Digital Investing; Econometric Analysis; Robo-Advisors; Micro-Investment Platforms; Impact Finance; Portfolio Democratization; Green Digital Ecosystems.},
        month = {June},
        }