Figure 5From: LEIA: Linguistic Embeddings for the Identification of AffectComparison of F1 score for each label on the out-of-domain datasets between LEIA and GPT models. Error bars represent confidence intervals computed using bootstrapping with replacementBack to article page