Model Scaling
A strategy for improving model capability by increasing the number of model parameters or the scale of training data. It is a method of expanding the capability boundary of a model, distinct from framework optimizations that do not change model weights.
Timeline (last 90 days)
In both static prediction and dynamic tasks, the framework's pass@1 advantage over stronger interventions (such as model scaling or increased inference compute) may reverse under a larger attempt budget.
When the attempt budget increases (e.g., pass@10, pass@100), the pass@k curves of stronger interventions such as model scaling or increased inference compute may overtake framework optimization.
All Facts (2)
In both static prediction and dynamic tasks, the framework's pass@1 advantage over stronger interventions (such as model scaling or increased inference compute) may reverse under a larger attempt budget.
50%UnverifiedWhen the attempt budget increases (e.g., pass@10, pass@100), the pass@k curves of stronger interventions such as model scaling or increased inference compute may overtake framework optimization.
50%