Edit model card

BENCHMARKS: AVG: 62.97 ARC: 60.58 HellaSwag: 76.03 MMLU: 55.8 TruthfulQA: 52.64 Winogrande: 77.83 GSM8K: 55.72

3b variant of MFANNv0.5

fine-tuned on the MFANN dataset which is still a work in progress, and is a chain-of-thought experiment carried out by me and me alone.

image/png

Downloads last month
2,301
Safetensors
Model size
2.78B params
Tensor type
F32
·

Dataset used to train netcat420/MFANN3bv0.4