Liquid Foundation Models: Our First Series of Generative AI Models

Benchmark Model Rank Results
arithmetic-reasoning-on-gsm8kLFM-40B Preview (5-shot)–Accuracy: 76.04
multi-task-language-understanding-on-mmluLFM-40B Preview (5-shot)–Average (%): 78.76
multi-task-language-understanding-on-mmlu-proLFM-40B Preview (5-shot)–Accuracy: 55.63
multiple-choice-question-answering-mcqa-on-arc-cLFM-40B Preview (25-shot)–Accuracy: 67.24
sentence-completion-on-hellaswagLFM-40B Preview (10-shot)–Accuracy: 82.07