[KongchangAI]
Benchmark

CMMLU

A Chinese multitask language understanding evaluation benchmark covering multiple subject domains, used to assess the capabilities of Chinese large models.

Timeline (last 90 days)

Nov 4

The evaluation set consists of a fixed 5,112 samples, mixing ARC-Challenge, ANLI-R3, and stratified subsets of MMLU-Pro and CMMLU.

Unverified50%
Nov 3

On the 16B MoE model, DepthGraft achieves a 7.0% absolute improvement on MMLU-Pro, 6.6% on CMMLU, and 6.4% on C3.

Unverified50%

All Facts (2)

Source Articles