Adding a self-critique and refinement step to curiosity-driven prompting improves GPT-4o-judged honesty and helpfulness scores on HONESET by 1.4% to 4.3% across ten LLMs.
Aligning ai with shared human values,
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
-
Self-Critique-Guided Curiosity Refinement: Enhancing Honesty and Helpfulness in Large Language Models via In-Context Learning
Adding a self-critique and refinement step to curiosity-driven prompting improves GPT-4o-judged honesty and helpfulness scores on HONESET by 1.4% to 4.3% across ten LLMs.