deliciouscat
/

Qwen3-8B-A1.4B-base-unhealed

Model card Files Files and versions

Qwen3-8B-A1.4B-base-unhealed

MoEfied Qwen3-8B-base, and ablated attn layer 16, 29 (by 0-based index)

Downloads last month: -

Safetensors

Model size

8B params

Tensor type

BF16

·

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support