Small and large LLMs often share the same optimal prompt, so a small model can select effective prompts for a larger model at lower cost.
We present the man- ually crafted prompt templates below and highlight the best-performed prompt template on the target model, GPT-3.5, in bold
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
-
S2LPP: Small-to-Large Prompt Prediction across LLMs
Small and large LLMs often share the same optimal prompt, so a small model can select effective prompts for a larger model at lower cost.