Yuhao Yao

[bugfix] fix prefill TBO crashes when attn_tp_size > 1

1

## Motivation Issues [#12757](https://github.com/sgl-project/sglang/issues/12757) and [#10863](https://github.com/sgl-project/sglang/issues/10863) indicate that the current TBO implementation still has several bugs. Although PRs such as [#11423](https://github.com/sgl-project/sglang/pull/11423) and [#13082](https://github.com/sgl-project/sglang/pull/13082) address part of the problem, some configurations...

Hi DeepSeek team, I noticed that `num_worst_tokens` is included in the arguments of dispatch, which makes sense for CUDA Graph support. However, it seems that the values written to `moe_recv_counter`...