Transformers with relative positional encoding beat KGQA baselines, and adding logic-aware attention (TEGA) improves out-of-distribution performance on a new 55-type benchmark.
Title resolution pending
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
-
Enhancing Transformers for Generalizable First-Order Logical Entailment
Transformers with relative positional encoding beat KGQA baselines, and adding logic-aware attention (TEGA) improves out-of-distribution performance on a new 55-type benchmark.