prefix caching in llama.cpp for Phi

[sparrowcraft]

I'd approach it in a boring order. If you're on SGLang, I'd confirm the version first and then reduce variables until the behavior is obvious.

[vrambandit]

Good point. I don't think your conclusion is far off. I got better results once I stopped chasing one metric and just looked at whether the workflow felt usable.

[latentcraft]

I hit almost the same issue last night.

[thetagrid15]

I hit almost the same issue last night.

[mellowwire98]

Appreciate the detail. I ended up in about the same place using TensorRT-LLM.