ChineseQA

Models answer Chinese trivia questions in a closed-book format modeled on TriviaQA.

In-house: DeepSeek never released the questions.

Made by DeepSeek Released Nov 2023 Trivia and open questions Tasks kept private Last in a launch post Nov 2023 , dropped after it saturated Last model added Nov 2023

Level: Everyday Most adults, without training
Scored: Checked automatically A program checks the result against an answer key, unit tests or the expected final state.
Answers: Short answer A number, word or short string matched to a key.

Model scored in lab cards

Launch post lists it, from 1 lab

DeepSeek LLM 67B leads with 85.1%

Best score per model as labs report them in their own cards. Labs pick their own settings, so these compare less cleanly than independent runs.

Bars Table

85.1 DeepSeek LLM 67B

1 DeepSeek LLM 67B 85.1%

#	Model	Score	Setting	Date	Source
1	DeepSeek LLM 67B	85.1%		Nov 2023	Model card

The best score passed 85% in Nov 2023

All the scores we have, one dot each, so a model tested at several settings appears more than once.

Reported by the lab in its model card

Labs dropped it from launch posts once the best score passed 85%

Launch posts from every lab that has listed it, one tick per post. Tall colored ticks include it, and short gray ones leave it out.

DeepSeek

Date	Launch post	Score	As the post writes it
29 Nov 2023	DeepSeek LLM 67B	85.1	ChineseQA (in-house)

DeepSeek LLM 67B leads with 85.1%

The best score passed 85% in Nov 2023

Labs dropped it from launch posts once the best score passed 85%

More in Knowledge