Surprise-driven continual learning absorbs planted lies about known facts
A model's own prior separates known-fact lies at AUC 0.78–0.83 but not legitimate updates, at AUC 0.53.
Stories per day · 30 days
Relatedcontinual-learning1data-poisoning1interpretability1leace1lora1paraphrase1sheaf-theory1transformers1
2 stories, newest first · one card per story
Stories tagged #counterfact
Sheaf method isolates paraphrase-stable content in transformers
A 20-dimensional content subspace beats LEACE by a mean 17.1 points on held-out CounterFact retrieval across eight architectures.