Prompt rankings for LLMs show moderate overall correlation but frequent changes in the top prompt under random seeds and limited subsets; a lower-confidence-bound selection strategy improves robustness.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
On the Stability of Prompt Ranking in Large Language Model Evaluation
Prompt rankings for LLMs show moderate overall correlation but frequent changes in the top prompt under random seeds and limited subsets; a lower-confidence-bound selection strategy improves robustness.