evalengine/decision-4b
Adapter vs Qwen/Qwen3.5-4B. evalengine/decision-4b is a LoRA adapter, not full weights: rank 8, on down_proj, gate_proj, in_proj_a, in_proj_b, in_proj_qkv, in_proj_z, k_proj, o_proj… in 32 layers (248 adapted matrices). It modifies Qwen/Qwen3.5-4B at load time.
1 derivatives checked
- evalengine/decision-4b-gguf: Adapter