Purpose
A measuring instrument, not a chatbot.
BarunLM is an English-centric base completion model. It is designed for architecture research, controlled evaluation, and local experimentation rather than instruction-following chat.
Open model
BarunLM 35M
A compact decoder-only base model for efficient language research and reproducible architecture experiments.
Open on Hugging FaceBarunLM 35M
Open weightsCompact base model
Capability per parameter.
BarunLM is a compact decoder-only base model created to study how architecture, data quality, and evaluation interact under hard compute limits.
- Parameters
- 35.07M
- Layers
- 12
- Context
- 2,048
- Training tokens
- 5.7B
Design
Local attention, then global context.
The twelve-layer network repeats a three-local, one-global attention pattern. This creates a useful laboratory for studying efficiency and information flow.
Release standard
Evidence travels with the artifact.
Weights, configuration, evaluation details, limitations, and usage guidance are treated as one release surface.