CE
Cerebras
Summary
Cerebras covering Large Language Models (LLMs), Model Scaling, Data Scaling, and Mixture-of-Experts (MoE). Notable guests include Junyang Lin (Leader of the Qwen LLM team, Alibaba Cloud).
1episodes
19total claims
12topics covered
1 episodes
Junyang Lin Scaling Model Size and Context Length Towards Intelligence
›Jan 13, 2026 · WITH Junyang Lin
Data scaling is considered more critical than model scaling, with the Qwen team's training data growing from 2T to 36T tokens, utilizing a multi-stage process that refines data quality over time. M...
Large Language Models (LLMs)Model ScalingData ScalingMixture-of-Experts (MoE)+14 more