04:00
2026-09-15
arxiv.org
large-language-models
Lexical Prompt Compression for Large Language Models: A Training-Free, Deterministic Pipeline with Empirical Pareto Analysis Across Eleven Task Categories
A training-free, CPU-only lexical prompt-compression pipeline reduced prompt tokens by a mean of 40.3% (sigma = 9.2) while holding BERTScore-F1 at 0.876 against original-prompt output, according to an…