Bilingual LMs ( L1 {es fr de pl tr ar zh} + L2 en ) trained on Cultura-X for L1 and FineWebEdu (L2)
Suchir Salhan
suchirsalhan
AI & ML interests
Multilinguality and Cognitively-Inspired AI. Tokenization, Pretraining, Interpretability & Alignment.
Organizations
models 226
suchirsalhan/qwen3.5-0.8b-playpen-rest
0.8B • Updated • 4
suchirsalhan/B-GPT-ar-fineweb-fluent
0.1B • Updated • 3
suchirsalhan/B-GPT-zh-fineweb-fluent
0.1B • Updated • 5
suchirsalhan/B-GPT-tr-fineweb-fluent
0.1B • Updated • 4
suchirsalhan/B-GPT-zh-fineweb-advanced
0.1B • Updated • 6
suchirsalhan/B-GPT-ar-fineweb-advanced
0.1B • Updated • 5
suchirsalhan/B-GPT-tr-fineweb-advanced
0.1B • Updated • 4
suchirsalhan/B-GPT-zh-fineweb-intermediate
0.1B • Updated • 5
suchirsalhan/B-GPT-ar-fineweb-intermediate
0.1B • Updated • 4
suchirsalhan/B-GPT-tr-fineweb-intermediate
0.1B • Updated • 8
datasets 11
suchirsalhan/rubric-adherence
Viewer • Updated • 61 • 23
suchirsalhan/celex-en-phoneme-sample
Viewer • Updated • 1.2k • 14 • 1
suchirsalhan/kidalign-llama-filterable
Viewer • Updated • 97.6k • 9
suchirsalhan/kidalign-llama-3.1-8B-Instruct
Updated • 6
suchirsalhan/babylm-detox
Viewer • Updated • 11.6M • 57
suchirsalhan/gptbert-tokenised
Updated • 17
suchirsalhan/Phonemized-UD
Viewer • Updated • 1.19M • 475
suchirsalhan/BabyLM-Pretokenised
Viewer • Updated • 1.64M • 21
suchirsalhan/MAO-CHILDES
Viewer • Updated • 3.81M • 26
suchirsalhan/CLiMP
Preview • Updated • 31 • 1