Pramana: Fine-Tuning Large Language Models for Epistemic Reasoning through Navya-Nyaya

Researchers present Pramana, a fine-tuning method that incorporates epistemic reasoning principles from Navya-Nyaya philosophy to improve large language models' knowledge validation and reasoning. The approach combines classical logical frameworks with modern deep learning to enhance how LLMs reason about evidence and knowledge claims.