Search papers, labs, and topics across Lattice.
NC State University
1
0
2
Large language models misclassify sentiment in French, favoring negative reviews, while encoder models in Japanese overlook critical negative feedback due to a positive bias.