Unique contextual benchmarks
GM Bench
Our first context bench is still in the works.
A context bench asks whether a model can keep the frame stable when the same object means different things in different tasks.
Unique contextual benchmarks
Our first context bench is still in the works.
A context bench asks whether a model can keep the frame stable when the same object means different things in different tasks.
.jpg){kind=link}