Picture for Leonidas Gee

Leonidas Gee

Code-Optimise: Self-Generated Preference Data for Correctness and Efficiency

Add code
Jun 18, 2024
Viaarxiv icon

Are Compressed Language Models Less Subgroup Robust?

Add code
Mar 26, 2024
Viaarxiv icon

Fast Vocabulary Transfer for Language Model Compression

Add code
Feb 15, 2024
Figure 1 for Fast Vocabulary Transfer for Language Model Compression
Figure 2 for Fast Vocabulary Transfer for Language Model Compression
Figure 3 for Fast Vocabulary Transfer for Language Model Compression
Figure 4 for Fast Vocabulary Transfer for Language Model Compression
Viaarxiv icon

Multi-Word Tokenization for Sequence Compression

Add code
Feb 15, 2024
Figure 1 for Multi-Word Tokenization for Sequence Compression
Figure 2 for Multi-Word Tokenization for Sequence Compression
Figure 3 for Multi-Word Tokenization for Sequence Compression
Figure 4 for Multi-Word Tokenization for Sequence Compression
Viaarxiv icon