Quiz · 5 questions
🧭 The Four Axes of Modern LLMs
Backbone, training, inference, product — the compass for everything else
Level 0Scout
0 XP0/41 lessons0/12 achievements
0/100 XP to next level100 XP to go0% complete
Quiz
01A new LLM is announced as '200B parameters with state-of-the-art reasoning'. What is the FIRST follow-up question?
Hint
Cost shape comes before benchmark shape.
02Which of these is a genuine *backbone architecture* difference (axis 1)?
Hint
Backbone = what is wired, not what is wrapped.
03DeepSeek-R1 and DeepSeek-V3 share the same MoE backbone (671B-A37B). Which axis explains the difference in their behavior?
Hint
If two models have identical config.json, the difference is somewhere else.
04A model is described as 'reasoning-enabled'. What does this most likely mean in 2026?
Hint
Two axes light up at once: training and inference.
05Two models advertise '70B parameters', but Model A feels twice as fast as Model B at the same provider price. The most likely explanation is:
Hint
If two models with the same advertised params feel different, ask 'same number, different shape' — start with active per token.
Comments 0
🔔 Reply notifications (sign in)Sign in — Please sign in to comment.
No comments yet — be the first.