exnivo/tinybrain-100m-instruct

TinyBrain-100M Instruct

TinyBrain-100M Instruct is a small chat/instruct model fine-tuned from exnivo/tinybrain-100m-base on exnivo/tinybrain-instruct.

Model details

• Parameters: ~103M
• Architecture: LLaMA-style causal transformer
• Tokenizer: custom TinyBrain byte-level BPE
• Vocab size: 24,000
• Base model: exnivo/tinybrain-100m-base
• SFT dataset: exnivo/tinybrain-instruct

Prompt format

<|bos|><|user|>
Your message here
<|end|>
<|assistant|>

Limitations

This is a very small model. It can hallucinate, make factual mistakes, fail at math, and give unreliable answers. Do not use it for medical, legal, financial, or safety-critical decisions.