Skip to content

Author

Dongsheng Wang

We have 1 of 2 papers

We haven’t gathered this author’s papers yet. Follow them and we’ll fetch their work.

Not the right person? Other researchers publish under this name.

Preprint Jul 2026

Request-Level Energy Attribution for Batched LLM Serving

The results show that token attribution is not a reliable proxy for marginal energy under batched execution, and that measured Shapley ground truth can calibrate low-cost request features toward fairer attribution.

Qi Luo, Kun Li, Ziwen Wang et al. · 0 citations