OpenBookQA

Measures: Four-option elementary science questions that need one core science fact from a small 'book' combined with broad common knowledge.

Leaderboard

Independent and third-party results only.

Leaderboard
Model	Accuracy	Provenance	Source	Evaluated
text-davinci-001 (InstructGPT) OpenAI	65.4%	Third party	GLaM: Efficient Scaling of Language Models with Mixture-of-Experts via Epoch AI	—
text-davinci-001 (InstructGPT) OpenAI	65.4%	Third party	Language Models are Few-Shot Learners via Epoch AI	—
text-davinci-001 (InstructGPT) OpenAI	58.8%	Third party	GLaM: Efficient Scaling of Language Models with Mixture-of-Experts via Epoch AI	—
text-davinci-001 (InstructGPT) OpenAI	58.8%	Third party	Language Models are Few-Shot Learners via Epoch AI	—
text-davinci-001 (InstructGPT) OpenAI	57.6%	Third party	GLaM: Efficient Scaling of Language Models with Mixture-of-Experts via Epoch AI	—
text-davinci-001 (InstructGPT) OpenAI	57.6%	Third party	Language Models are Few-Shot Learners via Epoch AI	—
text-davinci-001 (InstructGPT) OpenAI	57.6%	Third party	LLaMA: Open and Efficient Foundation Language Models via Epoch AI	—
Falcon-40B Instruct Technology Innovation Institute (TII)	56.6%	Third party	Llama 2: Open Foundation and Fine-Tuned Chat Models via Epoch AI	—
MPT-30B (Instruct, Chat) Databricks	52.0%	Third party	Llama 2: Open Foundation and Fine-Tuned Chat Models via Epoch AI	—
Falcon-7B Instruct Technology Innovation Institute (TII)	51.6%	Third party	Llama 2: Open Foundation and Fine-Tuned Chat Models via Epoch AI	—
MPT-7B (Instruct, Chat, StoryWriter) Databricks	51.4%	Third party	Llama 2: Open Foundation and Fine-Tuned Chat Models via Epoch AI	—