Test-time scaling should be tuned for wall-clock latency rather than token count; branches plus speculative decoding win under time budgets.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Faster and Better LLMs via Latency-Aware Test-Time Scaling
Test-time scaling should be tuned for wall-clock latency rather than token count; branches plus speculative decoding win under time budgets.