BSM-SM interference simulation

Dear experts,

I'm currently using the VPrime_NLO model to generate W' events in s-channel with the SM interference. However, I encountered a problem that I could not understand. I would appreciate it if you could give me some opinion.

I simulated the samples as follows:

W background (sm model)
p p > w+ w+ > e+ ve
p p > w- w- > e- ve~

interference (VPrime_NLO model)
p p > wp+ wp+ > e+ ve
p p > wp- wp- > e- ve~
p p > w+ w+ > e+ ve
p p > w- w- > e- ve~

Decay width of Wprime is calculated using MWp and coupling strength kL. Also, I exclusively simulated each process by using charged lepton pt cut to simulate up to the high mt region and merged in the last. Below is the information about pt cut range and corresponding cross sectoin.

W background

charge lepton pt cut[GeV] | cross section [pb]

 10 ~ 200 | 9682.1
 200 ~ 500 | 2.130 × 10^(−1)
 500 ~ 1000 | 5.172 × 10^(−3)
1000 ~ 1500 | 1.354 × 10^(−4)
1500 ~ inf | 9.123 × 10^(−6)

interference (MWp = 6 TeV , kL = 0.3 , kR = 0)

charged lepton pt cut [GeV] | cross section [pb]
  10 ~ 100 | 9499.4
 100 ~ 500 | 2.52
 500 ~ 2500 | 5.25×10^(−2)
2500 ~ 3500 | 2.92×10^(−7)
3500 ~ 4000 | 9.51×10^(−12)
4000 ~ inf | 4.96×10^(−13)

I compared the reconstructed mt distribution of the two processes on the generator level. The problem is that the interference sample shows lower yield than the W background on the high mt region where the constructive interference effect occurs.

For example, at MWp = 6 TeV and kL = 0.3, the yield of interference sample becomes lower than the W background starting from the 7 TeV on mt region.

I wonder what has caused issues in simulating the interference sample on the high mt region.