DUD decouples Attention and FFN contributions via noise-induced causal patching and trains a probe on the dual-stream restoration scores, reporting state-of-the-art hallucination detection.
late- layer
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
DUD: Decoupled Update Dynamics for Reliable Uncertainty Quantification in Large Language Models
DUD decouples Attention and FFN contributions via noise-induced causal patching and trains a probe on the dual-stream restoration scores, reporting state-of-the-art hallucination detection.