Pretrained lexical priors in language models persist despite explicit remapping rules, as shown by a Stroop paradigm where prior strength predicts interference and activation patching localizes the repair mechanism.
arXiv preprint arXiv:2602.17520 , year =
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Priors Persist Through Suppression: A Stroop Paradigm for Lexical Override
Pretrained lexical priors in language models persist despite explicit remapping rules, as shown by a Stroop paradigm where prior strength predicts interference and activation patching localizes the repair mechanism.