diff --git a/src/RobustModels.jl b/src/RobustModels.jl
index ac8e22b..d556390 100644
--- a/src/RobustModels.jl
+++ b/src/RobustModels.jl
@@ -191,8 +191,9 @@ abstract type RobustResp{T} <: ModResp end
 
 abstract type AbstractRegularizedPred{T} end
 
-Base.broadcastable(m::T) where {T<:AbstractEstimator} = Ref(m)
-Base.broadcastable(m::T) where {T<:LossFunction} = Ref(m)
+Base.broadcastable(m::AbstractEstimator) = Ref(m)
+Base.broadcastable(m::LossFunction) = Ref(m)
+Base.broadcastable(m::PenaltyFunction) = Ref(m)
 
 include("tools.jl")
 include("losses.jl")
diff --git a/src/penalties.jl b/src/penalties.jl
index d7ba1fc..b9560c0 100644
--- a/src/penalties.jl
+++ b/src/penalties.jl
@@ -97,7 +97,7 @@ function proximal!(p::SquaredL2Penalty{T}, out, x::AbstractVector{T}, step::T=on
 #    return broadcast!(/, out, x, 1 + p.λ * step)
     a = 1 / (1 + p.λ * step)
     @inbounds @simd for i in eachindex(out, x)
-        out[i] = x[i] > 0 ? x[i] * a : zero(T)
+        out[i] = (p.nonnegative && x[i] <= 0) ? zero(T) : x[i] * a
     end
     return out
 end
@@ -122,7 +122,16 @@ struct EuclideanPenalty{T<:AbstractFloat} <: PenaltyFunction{T}
 end
 cost(p::EuclideanPenalty{T}, x::AbstractVector{T}) where {T<:AbstractFloat} = p.λ * norm(x, 2)
 function proximal!(p::EuclideanPenalty{T}, out, x::AbstractVector{T}, step::T=one(T)) where {T<:AbstractFloat}
-    nn = p.nonnegative ? norm(broadcast!(max, out, x, 0), 2) : norm(x, 2)
+#    nn = p.nonnegative ? norm(broadcast!(max, out, x, 0), 2) : norm(x, 2)
+    if p.nonnegative
+        nn = zero(T)
+        @inbounds @simd for xi in x
+            nn += xi > 0 ? xi^2 : zero(T)
+        end
+        nn = sqrt(nn)
+    else
+        nn = norm(x, 2)
+    end
     return rmul!(copyto!(out, x), (1 - p.λ * step / max(p.λ * step, nn)))
 end