Model Scale

حجم النموذج

مقياس يشمل عدد معاملات النموذج وحجم بيانات تدريبه والحوسبة المستخدمة. الورقة أظهرت أن سلسلة التفكير لا تُفيد إلا عند حجم يتجاوز ~100 مليار معامل.

A measure encompassing the model's parameter count, training data size, and compute used. The paper showed chain-of-thought only helps beyond ~100B parameters.

Also translated asمقياس النموذج، سعة النموذج

First appears in this corpus in: Chain-of-Thought Prompting Elicits Reasoning in Large Language Models (2022)

Appears in these papers

Chain-of-Thought Prompting Elicits Reasoning in Large Language Models2022 in the sky ✦
Chain-of-Thought Prompting Elicits Reasoning in Large Language Models2022 in the sky ✦
Emergent Abilities of Large Language Models2022 in the sky ✦
Emergent Abilities of Large Language Models2022 in the sky ✦
Are Emergent Abilities of Large Language Models a Mirage?2023 in the sky ✦