On Emergent Capabilities and Model Merging
Fine-tuned checkpoints and adapters now fill public repositories, and the most common operation applied to these artifacts is model merging: arithmetic on their weights that assembles capabilities cheaply. We ask what this operation does to emergent capabilities: behaviors an artifact carries that were never an explicit training target. Studying two independent testbeds (activation oracles and…
We haven't written up this one. arXiv cs.AI has the full story — the link below goes straight to it.