HARP improves LLM accuracy by up to 5.16% at about 1.25x inference cost, by running a dropout-perturbed second forward pass only when token-level entropy is high.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
HARP: Hesitation-Aware Reframing in Transformer Inference Pass
HARP improves LLM accuracy by up to 5.16% at about 1.25x inference cost, by running a dropout-perturbed second forward pass only when token-level entropy is high.