Quiz · 5 questions
🔍 Reading Model Announcements
Five questions, one phrasebook, no hype
Level 0Scout
0 XP0/41 lessons0/12 achievements
0/100 XP to next level100 XP to go0% complete
Quiz
01A model is announced as '235B-A22B parameters'. What does this notation mean?
Hint
X total, A means active per token.
02An announcement headlines 'state-of-the-art on MMLU'. Why should you be cautious?
Hint
If everyone scores 87–92%, the differences are noise.
03A model can operate in both 'fast' and 'thinking' mode from the same checkpoint (Qwen3-style). What does this prove about the reasoning capability?
Hint
Same weights file ⇒ only inference can differ.
04Most open-weight LLMs (Llama, Qwen, Mistral, DeepSeek) are best described as:
Hint
If you have not read the actual license file, you do not know.
05When reading a model card and a system card, which statement is most accurate?
Hint
Two documents, two angles, both worth reading before deployment.
Comments 0
🔔 Reply notifications (sign in)Sign in — Please sign in to comment.
No comments yet — be the first.