Adaptively quantizing parts of an LLM's layers to FP4 can improve win rates and trading yields in latency-sensitive agent tasks, but the reported gains come from choosing the best compression level after seeing test results.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Win Fast or Lose Slow: Balancing Speed and Accuracy in Latency-Sensitive Decisions of LLMs
Adaptively quantizing parts of an LLM's layers to FP4 can improve win rates and trading yields in latency-sensitive agent tasks, but the reported gains come from choosing the best compression level after seeing test results.