Quickstart examples missing "residual connection" lines

[jamesbraza]

From https://arxiv.org/pdf/2406.04692 section 2.3:

Each layer comprises a set of n expert networks alongside a gating network and includes residual connections for improved gradient flow.

In the repo root's README.md, the quickstart images:

• https://github.com/togethercomputer/MoA/blob/main/assets/moa-explained.png
• https://github.com/togethercomputer/MoA/blob/main/assets/moa-3layer.png

Are missing the "skip connection" / "residual connection" where the input prompt is also passed into all layers of MoA.