Input reduction reduces the length of the input context before feeding to downstream LLMs.

The in-context learning (ICL) capability of pre-trained models based on the transformer architecture has received growing interest in recent years.

Thus, we propose LLaMo: L arge La nguage Model-based Mo lecular graph assistant, which is an end-to-end trained large molecular graph language model.

CNNs, MLPs, and quadratic problems, and find that SGD can perform on par with Adam on problems without block heterogeneity, but performs worse than Adam when the heterogeneity exists.