A multi-task GPT-4 model is said to beat single-task GPT-4, GPT-3, BERT, and Bi-LSTM on classification and summarization, but the experimental evidence is not reported.
Effective design space exploration of gradient nanostructured materials using active learning based surrogate models,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2024 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Optimizing Multi-Task Learning for Enhanced Performance in Large Language Models
A multi-task GPT-4 model is said to beat single-task GPT-4, GPT-3, BERT, and Bi-LSTM on classification and summarization, but the experimental evidence is not reported.