FewCLUE: TNEWS (Short News Classification)

FewCLUE's short news classification task: sort a Chinese Toutiao headline into one of 15 categories, learned from 240 labelled training examples.

unassessed

This page is a discovery lead. Nobody has yet assessed it against the catalogue contract, so it carries no disposition. Absence of evidence here is not evidence of staleness.
Categoryreasoning
Subcategoryshort Chinese news headline topic classification (15-way), few-shot
Page statusunknown
Metricaccuracy
Directionhigher_is_better
Unit%
Dataset size2010
PublisherCLUE team

What it measures

A short Chinese news headline from the Toutiao platform; the model classifies it into one of 15 topic categories such as finance, sports or education, learned few-shot from 240 labelled training headlines (16 per category).

Task format

Fifteen-way topic classification, graded on the single correct category; evaluated from a 240-example few-shot training split (16 per class), one of five parallel splits FewCLUE provides for this task.

Models reporting this benchmark

No model card in ModelSpec reports this benchmark yet.

Data

This page as JSON · Edit on GitHub