Function glm_hc_post 

pub fn glm_hc_post(
    gpu: &dyn GpuBackend,
    kernel: KernelHandle,
    block_out: DevicePtr,
    residual: DevicePtr,
    post: DevicePtr,
    comb: DevicePtr,
    out: DevicePtr,
    num_tokens: u32,
    hidden_size: u32,
    hc_mult: u32,
    stream: u64,
) -> Result<()>

2026-09-25: out[j] = post[j] * block_out + Σ_i comb[i][j] * residual[i] for each token. out may alias residual: the kernel reads every residual stream of an element before it writes that element.