You are developing an NLP model to perform sentiment analysis on customer reviews using a bidirectional LSTM network. Initially, the model struggles to differentiate between slightly negative and slightly positive sentiments$.$ Which technique could potentially improve the model's performance in distinguishing these nuances?

Increasing the number of LSTM layers in the network.

Reducing the complexity of the model by using fewer parameters.

Incorporating word embeddings like GloVe or Word$2$Vec that capture semantic meanings.

Switching from LSTM to a simpler logistic regression model.