$
$
Nate Krasner

about

I'm a PhD candidate at George Mason University researching multilingual alignment in large language models — getting LLMs to share knowledge between languages instead of behaving like a different model per language. I'm also a data scientist at The Asia Group. I work mostly in Python / PyTorch and care about making AI systems work as well for the rest of the world's languages as they do for English.

LLMs have many "brains" inside of them.

They "think" differently depending on what language you speak.

English
Mandarin
Arabic
Hindi
Japanese
Aligned Brain

I create alignment methods to bring them together, sharing knowledge between languages.

papers

2025

Cross-Lingual Representation Alignment Through Contrastive Image-Caption Tuning

Nathaniel Krasner, Nicholas Lanuzo, Antonios Anastasopoulos. ACL 2025 (Short Papers).

[acl] [arxiv] [pdf]

2025

Machine Translation Metrics for Indigenous Languages Using Fine-tuned Semantic Embeddings

Nathaniel Krasner, Justin Vasselli, Belu Ticona, Antonios Anastasopoulos, Chi-Kiu Lo. AmericasNLP 2025.

[acl] [pdf]

2025

Machine Translation Using Grammar Materials for LLM Post-Correction

Jonathan Hus, Antonios Anastasopoulos, Nathaniel Krasner. AmericasNLP 2025.

[acl] [pdf]

2022

Revisiting the Effects of Leakage on Dependency Parsing

Nathaniel Krasner, Miriam Wanner, Antonios Anastasopoulos. Findings of ACL 2022.

[acl] [arxiv]

projects

2024

JSON-Flow

A library for easily using LLMs in arbitrary JSON->JSON functions.

[github]

2022

Wordle Bot

A python program which solves Wordle puzzles.

[github]

contact

Drop a line — questions, collaborations, or just to say hi.