New Paper Explains Why Generative Models Need Diversity and How to Post-Train for the Right Amount
MartinKlissarov · x · 2026-09-11
A new paper by AntChen tackles a conceptual question: why should generative models be diverse at all — if there's an obviously best answer, why put probability on worse ones?
The paper offers a perspective on what diversity is actually for, and proposes a post-training algorithm that calibrates models to retain the right amount of diversity rather than collapsing onto a single mode. The author walks through the argument in a 12-tweet thread.
More from Research
- NYU mathematician says OpenAI pushed him to drop co-author's name after private call — eyishazyer · 2026-09-11
- NYU mathematician says OpenAI pushed to scrub collaborator from Navier-Stokes proof announcement — eyishazyer · 2026-09-11
- Arch breakdown: dropping convs for Muon optimizer, 3x3 pixel unshuffle for vision — stochasticchasm · 2026-09-11
- The 'Trickle Test': a new eval measuring whether user models disclose information gradually like humans — gharik · 2026-09-11
- Humans& Releases Persimmon, First Large-Scale Model Simulating How People Talk — niloofar_mire · 2026-09-11
- Yacine: RL can run in real time — almost nobody is seriously trying — yacineMTB · 2026-09-11