Skip to main content

🏢 Department of Computer Science, University of Chicago

Do LLMs dream of elephants (when told not to)? Latent concept association and associative memory in transformers
·2914 words·14 mins· loading · loading
Natural Language Processing Large Language Models 🏢 Department of Computer Science, University of Chicago
LLMs’ fact retrieval is easily manipulated by context, highlighting their associative memory behavior; this paper studies this with transformers, showing how self-attention and value matrices support …
Contextual Active Model Selection
·2539 words·12 mins· loading · loading
Machine Learning Active Learning 🏢 Department of Computer Science, University of Chicago
CAMS, a novel contextual active model selection algorithm, minimizes labeling costs by strategically selecting pre-trained models and querying labels for data points, achieving significant improvement…