Queueing-Aware Optimization of Reasoning Tokens for Accuracy-Latency Trade-offs in LLM Servers.

Emre Ozbas, Melih Bastopcu

Browse the full ISIT paper archive.