List your company

Language Modeling on Specialized Corpora (MIMIC-III, FinGPT, Asylex)

8.25Perplexity (MIMIC-III)Nirvana
8.20888.48698.7659.0431Oct 30, 2025

Evaluation results

MethodLinks
NirvanaFine-tuning=3 epochs, Model Scale=1.3B2025.108.257.887.227.78
Gated DeltaNetFine-tuning=3 epochs, Model Scale=1.3B2025.109.029.728.899.21
Mamba2Fine-tuning=3 epochs, Model Scale=1.3B2025.109.139.979.079.39
Nirvana-noTriggerFine-tuning=3 epochs, Model Scale=1.3B2025.109.199.878.849.3
SambaFine-tuning=3 epochs, Model Scale=1.3B2025.109.279.58.749.17
Transformer++Fine-tuning=3 epochs, Model Scale=1.3B2025.109.2810.78.829.6