Can I ask what your use case is for this?
I don’t think nanite translucency is a priority because translucent materials don’t really get the same benefits from Nanite. They can’t occlusion cull anything behind them and they’re rarely high enough in triangle count to get any benefit from streaming in only visible clusters, deferred shading evaluation, or software rasterization.
The fact that they’re all currently done separately in a forward pass probably also makes it complicated to do it in a performant way.