Model Overview

2B_or_not_2B is a 2.5 billion parameter language model developed by SicariusSicariiStuff, fine-tuned from an original Google model. The model's name is a playful nod to invisietch and Shakespeare. It was fine-tuned on a laptop with a 4090 16GB GPU, taking approximately 4 hours.

Key Characteristics

• Parameter Count: 2.5 billion parameters.
• Censorship Level: Rated at 7.9 out of 10 (low to very low censorship), making it suitable for less restricted applications.
• Quantizations: Available in multiple quantized formats, including GGUFs (Static, iMatrix_GGUF-bartowski, iMatrix_GGUF-mradermacher), EXL2 (8.0-BIT down to 4.0-BIT), FP8, and Mobile (ARM) Q4_0_X_X.

Benchmarks

The model reports the following benchmark scores:

• Average: 6.55
• IFEval (0-Shot): 20.62
• BBH (3-Shot): 7.68
• MATH Lvl 5 (4-Shot): 1.74
• GPQA (0-shot): 0.00
• MuSR (0-shot): 4.85
• MMLU-PRO (5-shot): 4.43

Use Cases

This model is particularly suited for applications requiring a language model with a low censorship threshold. Its availability in various quantized formats makes it adaptable for deployment on different hardware, including mobile devices.