Research4d ago
Pramana Fine‑Tunes LLMs with Ancient Logic
Pramana introduces a 6‑phase reasoning framework. By fine‑tuning LLMs on the 2,500‑year‑old Navya‑Nyaya logic, the model learns to ground claims in evidence, countering the 65% performance drop seen when irrelevant context is added. Practitioners in high‑stakes domains can adopt Pramana to reduce hallucinations and improve justification reliability and enhance transparency.