epfl-dlab/llm-latent-language

Repo accompanying our paper "Do Llamas Work in English? On the Latent Language of Multilingual Transformers".

36
/ 100
Emerging

This tool helps researchers and practitioners understand how multilingual large language models process and translate text between different languages. By analyzing the 'latent language' within these models, you can gain insights into their internal workings. You provide text in one language, and the tool helps visualize how the model transforms it to understand or translate into another language. This is ideal for computational linguists, NLP researchers, and machine translation specialists.

No commits in the last 6 months.

Use this if you need to analyze the internal representations of multilingual large language models during translation or text completion tasks to better understand their behavior.

Not ideal if you are looking for a tool to simply perform translations or generate text, rather than to deeply investigate model internals.

computational-linguistics natural-language-processing machine-translation-research model-interpretability multilingual-NLP
No License Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 9 / 25
Maturity 8 / 25
Community 19 / 25

How are scores calculated?

Stars

80

Forks

18

Language

Jupyter Notebook

License

Last pushed

Mar 11, 2024

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/transformers/epfl-dlab/llm-latent-language"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.