Nirav-Madhani/

smol-prompt-template

	Experimental Dataset: Stateful Prompting with Two-Pass Diffing




	
	1. The Landscape & The Gap

Current advancements in Large Language Model (LLM) adaptation typically focus on three specific axes:

Model Weights (What the model knows): Fine-tuning (SFT), LoRA, Adapter layers. Retrieval (What the model sees): RAG, Vector Stores, GraphRAG. Optimization (How the model behaves): RLHF, DPO, Chain-of-Thought (CoT).

The Missing Axis: There is significantly less research dedicated to… See the full description on the dataset page: https://huggingface.co/datasets/Nirav-Madhani/smol-prompt-template.