LLM Model Collapse

Model Collapse is a degenerative process in recursive training cycles where successive generations of model’s lose statistical fidelity due to cumulative discretization errors.