Presents a byte-native LLM with bespoke tokenizer achieving 69-98% accuracy on malware family and architecture classification from raw bytes.
likely”, “unlike
4 Pith papers cite this work. Polarity classification is still indexing.
4
Pith papers citing it
representative citing papers
LLMs can compose surface-form tokens from base embeddings plus learned transformation vectors, freeing 10-40% of vocabulary slots while expanding coverage and preserving downstream performance across five languages.
SemBridge initializes target tokens in sparse encoders via linear combinations of semantically related source tokens selected through multilingual dense bridge embeddings, yielding better zero-shot and fine-tuned retrieval across languages.
Replacing tokens, freezing the corresponding embeddings, and tuning the rest of the model improves NLU performance on low-resource languages compared to full fine-tuning.
citing papers explorer
-
Large Byte Model: Teaching Language Models About Compiled Code
Presents a byte-native LLM with bespoke tokenizer achieving 69-98% accuracy on malware family and architecture classification from raw bytes.
-
Vocab Diet: Reshaping the Vocabulary of LLMs via Vector Arithmetic
LLMs can compose surface-form tokens from base embeddings plus learned transformation vectors, freeing 10-40% of vocabulary slots while expanding coverage and preserving downstream performance across five languages.
-
SemBridge: Language Transfer in Sparse Encoders via Multilingual Semantic Bridges
SemBridge initializes target tokens in sparse encoders via linear combinations of semantically related source tokens selected through multilingual dense bridge embeddings, yielding better zero-shot and fine-tuned retrieval across languages.
-
Modular Monolingual Adaptation using Pretrained Language Models
Replacing tokens, freezing the corresponding embeddings, and tuning the rest of the model improves NLU performance on low-resource languages compared to full fine-tuning.