Direct preference alignment losses can be expressed as logical programs over model predictions, yielding an organized landscape of billions of definable losses and a route to new variants.
However, the semantics of the resulting formulas are less transparent and often hidden in the weights
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Understanding the Logic of Direct Preference Alignment through Logic
Direct preference alignment losses can be expressed as logical programs over model predictions, yielding an organized landscape of billions of definable losses and a route to new variants.