🇮🇳 India State Debt LLM (Chain-of-Thought Fiscal Reasoning Model)

A domain-specific Large Language Model fine-tuned on official Reserve Bank of India (RBI) and State Budget financial data for analyzing public liabilities, Debt-to-GSDP percentages, per-capita debt, national rankings, and multi-year growth trajectories (2020–2026) across 31 Indian States and Union Territories.

Usage with Hugging Face transformers

from transformers import AutoTokenizer, AutoModelForCausalLM, pipeline

model_id = "your-hf-username/india-state-debt-llm"

tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id)

generator = pipeline("text-generation", model=model, tokenizer=tokenizer)

prompt = "Instruction: Display Karnataka debt growth from 2020-2026\nResponse:\n<thought>\n"
response = generator(prompt, max_new_tokens=300, temperature=0.7)

print(response[0]["generated_text"])

Response Output Structure

The model generates responses structured into 3 explicit Chain-of-Thought stages:

• <thought>: Step-by-step query parsing, record identification, and fiscal metrics retrieval.
• <analysis>: Structured financial metrics, rank, per-capita figures, and category indicators.
• <answer>: Executive summary answer.