Many extensions of the Standard Model include neutral weakly-coupled particles that can be long- lived. These long-lived particles occur in many models, included gauge-mediated extensions of the Minimal Supersymmetric Model (MSSM), MSSM with R-parity violation, inelastic dark matter and the Hidden Valley scenario. Results are presented on the ATLAS searches at the LHC for possible rare Higgs boson decays to pair of neutral, long-lived hidden-sector particles that lead to final states containing collimated lepton jets or fermion anti-fermion pairs. No excess of events above the expected background has been observed on data collected in 2011 at a center of mass energy of 7 TeV and limits on the cross sections are set. of the leptons increases with d mass, but decreases with d. In this model, d = 0.0, 0.1, and 0.3, and d masses (m ) of 150, 300, and 500 MeV are used. For m = 150 MeV, the d is below the + threshold and can only de- cay to electrons. With m 300 MeV, the d decays to electron and muon pairs. Additionally, for m = 500 MeV, 20% of the decays produce pion pairs. This analysis con- siders LJs in three signatures: single muon-jets with four or more muons, pairs of muon-jets each with two or more muons, and pairs of electron-jets each with two or more electrons. Events containing electron-jets (EJs) were selected using single-electron triggers with an online pT threshold of 20 or 22 GeV. To ensure proper modelling of the trigger ac- ceptance, events were required to contain at least one re- constructed electron with pT > 35 GeV, above which the trigger e ciency is constant. The reconstructed electron was required to match an electron reconstructed above the pT threshold in the trigger system with a separation R less than 0.2. The EJ candidates were built from elec- tromagnetic (EM) clusters. At least two tracks from the primary vertex (PV) with pT > 10 GeV were required to have R < 0.1 of the cluster position in the second sam- pling layer of the calorimeter. Additional requirements were made on the number of hits along the track in the silicon pixel and microstrip detectors to suppress back- grounds from photon conversions. The analysis required two LJ candidates in each event, with one cluster match- ing the electron reconstructed in the trigger system. The invariant mass of the two highest-pT tracks associated with each EJ had to be less than 2 GeV. To reduce the back- ground coming from multi-jet events, the electron cluster concentration (R 2), defined as the ratio of total energy in 3 7 cells to the total energy in 7 7 cells in , in the second sampling layer of the EM calorimeter (ECAL) was used, together with the electron cluster lateral shower width and number of TRT high threshold hits. A scaled