Harnessing Explanations: LLM-to-LM Interpreter for Enhanced Text-Attributed Graph Representation Learning

2024-01-16 · ICLR 2024 poster · anchor

Findings

IC-996GPT-3.5 achieves 73.5% zero-shot accuracy on OGBN-ARXIV 40-class node classification and 73.56% on TAPe-ARXIV23, a dataset of papers published after its knowledge cutoff
IC-997LLaMA-2-13B-Chat achieves 44.23% zero-shot accuracy on OGBN-ARXIV, substantially below GPT-3.5's 73.5%
IC-998GPT-3.5's zero-shot accuracy on OGBN-ARXIV depends on the position of the title relative to the abstract in the prompt: 0.720 when abstract precedes title, 0.695 when title precedes abstract