Multiple Choice Qa
Multiple Choice Question Answering (MCQA) is a subtask of Natural Language Processing that requires the model to predict the best answer to a given question based on the provided candidate options and supporting context. This task aims to evaluate the model's comprehension and reasoning abilities and has a wide range of application values, such as intelligent education, online assessment, and knowledge retrieval, etc.
Chinchilla (few-shot, k=5)
Chinchilla (few-shot, k=5)
Chinchilla (few-shot, k=5)
Chinchilla (few-shot, k=5)
Chinchilla (few-shot, k=5)
Chinchilla (few-shot, k=5)
GAL 120B (zero-shot)
GAL 120B (zero-shot)
Gopher (few-shot, k=5)
Meditron-70B (CoT + SC)
CamemBERT
Med-PaLM 2 (ER)
GAL 30B (zero-shot)
Gopher (few-shot, k=5)
GAL 120B (zero-shot)
Chinchilla (few-shot, k=5)
GAL 120B (zero-shot)
Chinchilla (few-shot, k=5)