Function launch_k3_latent_moe_experts 

pub fn launch_k3_latent_moe_experts(
    gpu: &dyn GpuBackend,
    kernels: &K3MoeGemmKernels,
    packed: &[(String, QuantizedWeight)],
    latent: &[f32],
    ids: &[usize],
    mix_w: &[f32],
    cfg: &LatentMoeConfig,
    stream: u64,
) -> Result<Vec<f32>>

Packed w1/w2/w3 SiTU mix. Same contract as atlas_core::kimi_k3::mix_routed_experts.