Bounds on Partial Preparation Noncontextuality

We can examine the extent to which preparation contextuality is required to reproduce the predictions of quantum mechanics by considering models that are a convex mixture of a noncontextual model and a contextual model. The maximum weight that can be given to the noncontextual portion of such a model can be used to quantify the preparation contextuality required to explain quantum experiments. In this talk I will discuss bounds on this quantity from 2-bit parity-oblivious multiplexing and an extension to arbitrary angles. From this argument we can bound the overlap of two representations of the maximally mixed state.