GestaltLabs/Jeff-1
Adapter vs Qwen/Qwen3-4B-Instruct-2507. GestaltLabs/Jeff-1 is a LoRA adapter, not full weights: rank 16, on k_proj, o_proj, q_proj, v_proj in 36 layers (144 adapted matrices). It modifies Qwen/Qwen3-4B-Instruct-2507 at load time.
Adapter vs Qwen/Qwen3-4B-Instruct-2507. GestaltLabs/Jeff-1 is a LoRA adapter, not full weights: rank 16, on k_proj, o_proj, q_proj, v_proj in 36 layers (144 adapted matrices). It modifies Qwen/Qwen3-4B-Instruct-2507 at load time.