Text this: Advancing Spike Sorting Through Gradient‐Based Preprocessing and Nonlinear Reduction With Agglomerative Clustering