Symmetric Lanczos approximations: error (column index out of bounds)

I'm trying to calculate symmetric Lanczos approximations to exp(A)e_j (A is a 2361x2361 matrix, e_j is the j'th unit vector for j=1:2361) but can't get my code to work (I get the error "error: A(I,J): column index out of bounds; value 28 out of bound 1"). I think this is due to "lucky breakdown" but I still need the code to output an approximation for ALL j for use later on. My code:

 % code
end
n = length(A);
for m = 1:n
    v = eye(n)(:,m);
    V = v/norm(v);
    for j = 1:28
        w = A*V(:,j);
        for i = max(1,j-1):j
            T(i,j) = V(:,i)'*w;
            w = w - T(i,j)*V(:,i);
        end
        T(j+1,j) = norm(w);
        if T(j+1,j) < j*eps
           break;
        end
           V(:,j+1) = w/T(j+1,j);
    end
Y(:,m) = V(:,1:28)*(expm(T(1:28,1:28))*(norm(v)*eye(28,1)));
end

Does anyone have any ideas of how I could fix this? Thanks in advance.