PromptSET is a benchmark of 11,469 questions with nine LLM-generated rephrasings each, and current classifiers and self-evaluation methods predict prompt answerability poorly, especially on multi-hop questions.
In: Proceedings of the 31st ACM International Conference on Information & Knowledge Management
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
-
Benchmarking Prompt Sensitivity in Large Language Models
PromptSET is a benchmark of 11,469 questions with nine LLM-generated rephrasings each, and current classifiers and self-evaluation methods predict prompt answerability poorly, especially on multi-hop questions.