COPAL-ID (Choice of Plausible Alternatives — Local Nuances, Indonesia)

Two 559-item Indonesian COPA-style tests of Jakartan local terms, culture, and language, written from scratch in standard and colloquial Indonesian.

Also known as: COPAL-ID, COPAL, copal_id_standard, copal_id_colloquial

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.

Category	reasoning
Subcategory	Indonesian COPA-style causal commonsense with Jakartan local nuances
Page status	active
Metric	accuracy
Direction	higher_is_better
Unit	%
Dataset size	559
Dataset licence	CC-BY-SA-4.0 (Hugging Face card and README badge). The GitHub LICENSE file also opens with an MIT License line before the CC BY-SA 4.0 text.
Publisher	MBZUAI, with independent collaborators

What it measures

COPAL-ID is a two-choice causal commonsense task in Indonesian. The model reads one premise and a cause or effect cue, then picks the more plausible of two alternatives. Items were written from scratch by Jakartan natives, not translated from English COPA. They target local terminology, culture, and language phenomena that XCOPA-ID does not stress. Each item exists in standard Indonesian and in Jakartan colloquial Indonesian.

Task format

Two-choice classification: Indonesian premise plus cause/effect cue and two alternatives. lm-eval verbalizes the cue as "karena" (cause) or "maka" (effect) after the premise. No free-text explanation is required.

Models reporting this benchmark

No model card in ModelSpec reports this benchmark yet.

Data

This page as JSON · Edit on GitHub