Teaching LLMs to See Graphs: Unifying Text and Structural Reasoning
Open paper
Benchmark
Model
Rank
Results
molecule-captioning-on-chebi-20
GTLM (Llama-3.1-8B)
–
BLEU-2: 57.2
BLEU-4: 48.2
METEOR: 60.2
ROUGE-L: 58.3
molecule-captioning-on-chebi-20
GTLM (Llama-3.2-1B)
–
BLEU-2: 53.4
BLEU-4: 43.7
METEOR: 56.7
ROUGE-L: 55.1
Rank counts only results with a code link.