Suhas Kotha's repositories
understanding-forgetting
Understanding Catastrophic Forgetting in Language Models via Implicit Inference
verify-input
Provably Bounding Neural Network Preimages
predicting-emergence
Can you predict a larger model's predictions from smaller models?
purple-problem
You can't stop a language model from saying purple š¤·
highest_likelihood
Greedy decoding is not greedy enough
Language:Python000
lm-data-challenge
Can you find a good data ordering or data selection method?
Language:Python000