Update feature README.md

Hamiltonian_features/README.md

@@ -19,12 +19,6 @@ Number of qubits $n$
 
 Log FCI Size $\log_{10} \left( {N \choose N_{\uparrow}} {N \choose N_{\downarrow}}\right)$
 
-One-norm
-
-$$
-        \lambda(H) = \sum_{ij} |h_{ij}^{(1)}| + \frac{1}{2} \sum_{\ell = 1}^L |\lambda_\ell | \left( \sum_{pq} |g_{pq}^{(\ell)}| \right)^2
-$$
-
 Rank $L$
 
 Eigenvalues { $\lambda_\ell$ }
@@ -33,6 +27,12 @@ Double factorization eigenvalue gap $|\lambda_1 - \lambda_2|$
 
  $G(H) = (V,E)$ where $V = [n]$ for an $n$-qubit Hamiltonian $H$ where the edge set contains hyperedges $e_i = (i_1,...,i_{k(i)}) \in E$ where $i_1, ..., i_{k(i)} \in V$ are all those qubits that are being acted upon by non-identity single qubit Pauli operators. The graph has edge weights $w(e) = h_e$ where $h_e$ is the coefficient of Pauli string $e \in E$ where $H = \sum_{e \in E} h_e P_e$. We take statistics (max, min, mean, std. dev.) on edge order (Pauli weight), vertex degree, and edge weights.
 
+One-norm
+
+$$
+        \lambda(H) = \sum_{e \in E} |h_e|
+$$
+
 Number of Pauli Strings | $E$ |
 
 Edge Order $\mathrm{ord}(e_i) = k(i)$