Value preferences inferred from short-form LLM responses correlate only weakly (r around 0.05 to 0.25) with preferences inferred from long-form arguments, and alignment gives only modest consistency gains.
The steps for assigning the score are provided below:
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
-
Do Language Models Think Consistently? A Study of Value Preferences Across Varying Response Lengths
Value preferences inferred from short-form LLM responses correlate only weakly (r around 0.05 to 0.25) with preferences inferred from long-form arguments, and alignment gives only modest consistency gains.