A two-stage automatic prompting pipeline (error correction followed by guidance) reduces the accuracy loss large language models suffer when input questions contain typos, reordered words, or irrelevant extra information.
Large language models are human-level prompt engineers,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
Robustness of Prompting: Enhancing Robustness of Large Language Models Against Prompting Attacks
A two-stage automatic prompting pipeline (error correction followed by guidance) reduces the accuracy loss large language models suffer when input questions contain typos, reordered words, or irrelevant extra information.