Motivation

In the implementation of grouped topk in MoE layer, the scores of masked groups are set to 0, which may leads to select incorrect experts in certain scenarios.

Modifications

In SGLang, the mask scores are set to 0. This configuration may result in the selection of experts in masked groups if the scores in the unmasked groups are negative. This behavior can lead to incorrect or suboptimal selections in certain scenarios.

Checklist

• Format your code according to the Code Formatting with Pre-Commit.
• Add unit tests as outlined in the Running Unit Tests.
• Update documentation / docstrings / example tutorials as needed, according to Writing Documentation.
• Provide throughput / latency benchmark results and accuracy evaluation results as needed, according to Benchmark and Profiling and Accuracy Results.
• For reviewers: If you haven't made any contributions to this PR and are only assisting with merging the main branch, please remove yourself as a co-author when merging the PR.
• Please feel free to join our Slack channel at https://slack.sglang.ai to discuss your PR.