Matthew L Leavitt
4 papers · 2021–2025 · 2 venues · across top CS/AI venues
Authors are matched by name, so researchers who share a name ("Yang Liu", "Wei Wang") appear as a single profile and their paper counts, venues and achievements are combined. How identity is handled.
Achievements
Jump to papers ↓
🌍
Conference Polyglot · Bronze
🌉
Interdisciplinary Bridge · Bronze
🗺️
Taxonomy Completionist
(10)
Top co-authors
Research topics
Keywords
image classification
(1)
vision transformer
(1)
masked language model
(1)
phase transition
(1)
convolutional neural network
(1)
deep neural network
(1)
downstream task
(1)
training dynamics
(1)
data pruning
(1)
simplicity bia
(1)
language model pretraining
(1)
reference model
(1)
test accuracy
(1)
syntax acquisition
(1)
gated self-attention
(1)
convolutional inductive bia
(1)
positional self-attention
(1)
text dataset
(1)
class selectivity
(1)
syntactic attention structure
(1)
Papers
Perplexed by Perplexity: Perplexity-Based Data Pruning With Small Reference Models
ICLR 2025
Sudden Drops in the Loss: Syntax Acquisition, Phase Transitions, and Simplicity Bias in MLMs
ICLR 2024
ConViT: Improving Vision Transformers with Soft Convolutional Inductive Biases
ICML 2021
Selectivity considered harmful: evaluating the causal impact of class selectivity in DNNs
ICLR 2021