Feature Selective Model Collapse in Diffusion Models: Total Replacement versus Fixed-Budget Training
Fixed-budget diffusion retraining loses only fragile features, unlike full replacement with synthetic data.
4 stories, newest first · one card per story
Fixed-budget diffusion retraining loses only fragile features, unlike full replacement with synthetic data.
Synthetic-text source attribution collapses after rewriting and does not reliably curb model degradation.

OpenAI contractors were fired after using AI tools to review and train ChatGPT, violating Mercor’s ban on LLM-assisted work.
Training LLM reviewers on model-generated reviews compresses rating distributions and diversity, a failure the authors name scientific-judgment collapse and mitigate with TrustReviewer.
Today 1.6K readers · 2.5K views · 779 reads · all time 189K views