Checklist

• If this is not a feature request but a general question, please start a discussion at https://github.com/sgl-project/sglang/discussions. Otherwise, it will be closed.
• Please use English. Otherwise, it will be closed.

Motivation

Currently, dllm algorithms like low confidence adopt greedy rather than sampling with temperature.
We need to support sampling after batching PR #14883 (comment) is ready.

Related resources

No response