This thesis investigates the behavior of jets in heavy-ion collisions with the aim of advancing our understanding of the quark–gluon plasma (QGP), an extremely hot and dense state of matter that existed shortly after the Big Bang. In heavy-ion collisions at the Large Hadron Collider (LHC), conditions of sufficiently high energy density are achieved to produce this deconfined phase of quarks and gluons. Jets, collimated sprays of hadrons originating from high-energy quarks and gluons, serve as powerful probes of the QGP. As they traverse the medium, jets exchange energy and momentum with it, leading to modifications of their internal structure, a phenomenon known as jet quenching. The first part of the thesis outlines the theoretical framework underlying these studies, including the Standard Model of particle physics, Quantum Chromodynamics (QCD), and the space–time evolution of heavy-ion collisions. The physics of jet production, reconstruction, and jet quenching mechanisms, such as elastic and inelastic parton–medium interactions, are discussed in detail. A central objective of this work is to study correlations between jet substructure observables as a means of disentangling competing physical processes that affect jets in the QGP. While many previous measurements focus on single observables, multidimensional correlations provide enhanced sensitivity to the interplay between different quenching mechanisms. A phenomenological framework is developed to quantify the sensitivity of correlated observables to medium-induced modifications. Using simulated events generated with the JEWEL Monte Carlo model, scenarios with and without jet quenching are compared. The Hellinger distance is employed to systematically quantify differences between multidimensional distributions, allowing an assessment of which combinations of jet observables are most sensitive to QGP effects. Experimentally, the thesis presents a measurement of correlations between the Soft Drop jet substructure observables θ₍g₎ (groomed jet radius) and z₍g₎ (momentum sharing fraction) using data recorded by the ALICE experiment during Run 2 of the LHC at √sₙₙ = 5.02 TeV. The data are corrected for detector and background effects through unfolding techniques. The results demonstrate that jets produced in Pb–Pb collisions are, on average, narrower than those in proton–proton collisions at a given energy. This narrowing effect is observed for both balanced and unbalanced splittings, indicating medium-induced modifications of the jet structure. Comparisons with theoretical models, including JEWEL, Hybrid, and JETSCAPE, reveal that JEWEL tends to overestimate the observed narrowing, while JETSCAPE provides a more accurate description. Studies within the Hybrid model indicate sensitivity of the jet width to elastic parton–medium interactions. The results suggest that wide jets are more strongly modified by the medium than narrow jets. A key experimental challenge in measuring such correlations is the limited statistical precision inherent to multidimensional analyses, as the required data volume increases rapidly with dimensionality. The Run 2 dataset enabled the first measurement of these correlations but also highlights the importance of larger datasets. The ongoing Run 3 data-taking period at the LHC offers significantly increased statistics, providing opportunities for higher-precision and higher-dimensional studies of jet–medium interactions. Overall, this work establishes jet substructure correlations as a powerful tool for probing the microscopic dynamics of the quark–gluon plasma.