Background: In the bilateral Nordic hamstring exercise (NHE), hamstrings operate at relatively short lengths, which may limit this exercise’s efficacy in hamstring injury prevention. Objectives: To examine knee flexion torque, and biceps femoris long head (BFlh) and semitendinosus (ST) high-density electromyography (HD-EMG) activity during unilateral and bilateral NHE performed with either neutral (NHE0) or 90° flexed (NHE90) hips. Methods: Exercises were performed on a novel device at eccentric 1-repetition maximum load defined for 90-15° knee range of motion. Torque and EMG signals normalised to maximal voluntary isometric activity were compared in different phases of the exercises with Statistical Parametric Mapping. Results: Lower EMG levels were observed in NHE90 than in NHE0, mainly in the second half of the movement. Knee flexor eccentric torque was higher in NHE90 than in NHE0 from the beginning until 87% of the bilateral movement, and over the entire unilateral movement. In NHE0, ST activity was higher compared to BFlh during the initial movement phase, but lower close to knee extension. Torque and EMG activity were generally similar in the bilateral and unilateral modes. Conclusion: If performed with neutral hips, NHE selectively activates BFlh near full knee extension. Performing NHE with hips flexed to 90° is preferable when higher passive torque and ST selectivity are targeted at a longer muscle length. Performing these exercises unilaterally could help to train each limb separately with a similar torque and EMG output to the bilateral conditions. Adaptations to these exercises should be examined.