In this paper, we use the associated production of top-quark pairs ( toverline{t} ) with a generic scalar boson (ϕ) at the LHC (pp → toverline{t}phi ) to explore the sensitivity of a large set of observables to the sign of the CP mixing angle (α), present in the coupling between the scalar boson and the top quarks. The mass of the scalar boson is set to mϕ = 125 GeV (the Standard Model Higgs boson mass) and its coupling to top-quarks is varied such that α = 0°, 22.5°, 45.0°, 67.5°, 90.0°, 135.0° and 180.0°. Dileptonic final states of the toverline{t}phi system are used (pp → bℓ+νℓ overline{b} ℓ− overline{nu} ℓb overline{b} ), where the scalar boson is expected to decay according to ϕ → boverline{b} . A new method to reconstruct the scalar mass, originally designed for the low mass regime is used, improving the resolution of the Higgs mass by roughly a factor of two. A full phenomenological analysis is performed using Standard Model (SM) background and signal events generated with MadGraph5_aMC@NLO, in turn reconstructed using a kinematical fit. The most sensitive CP-observables are selected to compute Confidence Level (CL) limits as a function of the sign of the top quark Yukawa couplings to the ϕ boson. We also explore the sensitivity to interference terms using differential distributions and angular asymmetries. Given the significant difference between the pure scalar (σ0+) and pure pseudo-scalar (σ0−) production cross section values, it is unlikely the toverline{t}phi channel alone will be sensitive to the sign of the CP-mixing angle or interference terms, even at the end of the LHC. Using the {b}_2^{toverline{t}phi } and {b}_4^{toverline{t}phi } variables, exclusion limits at 95% CL for the CP-even and CP-odd components of the top quark Yukawa couplings are expected to be set to overset{sim }{kappa } ∈ [-0.698,+0.698] and |κ| ∈ [0.878,1.04], respectively, at the end of the High Luminosity phase of the LHC (HL-LHC) by using the dileptonic decay channel alone.