Small Model Post-Training Increments Portable to Large Models
tuvllms · x · 2026-07-09
This post introduces a new paper demonstrating that post-training parameter deltas from smaller models can be directly grafted onto larger models without needing to retrain the larger model from scratch.
The author notes that in certain settings, these grafted large models can even outperform their smaller "tutor" models, exhibiting a form of weak-to-strong generalization achieved during inference.
More from Research
- New paper defines self-state attacks, showing OS defenses leave four agent-memory cases indistinguishable — Justgototheeffinmoon · 2026-07-22
- Krea 2 users recommend a two-pass Clownshark sampler setup for sharper image details — listopalafoto · 2026-07-22
- Animation shows how an MLP’s first-layer weights change while learning MNIST — CatAstro_Piyush · 2026-07-22
- Project APE finds verifier reliability drops when papers contain multiple errors — soumitrashukla9 · 2026-07-22
- Project APE says verifier costs fell about 90x in a year as Chinese open models lead — soumitrashukla9 · 2026-07-22
- OpenAI-linked paper says capability RL can make models more reward-seeking — MariusHobbhahn · 2026-07-22