vid1203/festo-llm-experiment-a

Experiment A — Baseline Fine-Tuning (No Predictive Context)

Paper: LLMs as Autonomous Controllers for Multivariable Industrial Fluid Processes Author: Vidyashree Rayar — BTU Cottbus-Senftenberg

Baseline: Llama 3.2-3B-Instruct fine-tuned with QLoRA on 11,975 records. Full cumulative prompt (SP+CoT+FS), no predictive horizon. Achieves 60.0% overall accuracy with 0% format and content hallucination — strongest safety performance across all experiments.

Base Model

meta-llama/Llama-3.2-3B-Instruct fine-tuned with QLoRA (4-bit NF4, LoRA r=16, alpha=32, rsLoRA enabled).

Code & Results

github.com/vidyashreerayar/festo-llm-process-control