Biswajit Mishra
Biswajit Mishra has 2 papers in our library, in 2024–2026, most of them on Language models & pretraining, Semitic languages; often with Gurpreet Gosal, Rituraj Joshi, Abdelrahman Sadallah.
No OpenAlex profile matched yet.
2 works in our library · 21 citations in our library · ⭐ Paper Star rank #4646
Topics
Works by year
2020s
Papers
- Jais 2: A Family of Arabic-Centric Open Large Language Models (2026 · 14 citations)
Most large language models are trained mostly on English. Arabic, with its complex morphology, diglossia and many regional dialects, is poorly covered, so general multilingual models handle it only partly. This is especially true for dialects and culturally specific content. The Jais team (MBZUAI, Cerebras… - Bilingual Adaptation of Monolingual Foundation Models (2024 · 7 citations)
English-dominant open LLMs such as Llama 2 handle other languages poorly. Training a bilingual model from scratch is costly and tends to give weaker English. Adapting an existing model brings its own problems: it can forget English, and its English-centric tokenizer splits foreign text into many pieces. The authors…
Co-authors here
- Gurpreet Gosal ×2
- Rituraj Joshi ×2
- Abdelrahman Sadallah
- Abed Alhakim Freihat
- Bokang Jia
- Etienne Goffinet
- Joel Hestness
- Neha Sengupta
- Sunil Kumar Sahu
- Zhiming Chen
Cited by
This page in Markdown → https://md.research.elamur.ai/free/en/authors/biswajit-mishra/