Dataset · Hugging Face

DGurgurov/CLaS-Bench

Name: CLaS-Bench
Creator: DGurgurov
License: cc-by-4.0
Keywords: task_categories:text-generation, task_categories:translation, language:bo, language:mt, language:it, language:es, language:de, language:ja, language:ar, language:zh, language:nl, language:fr, language:pt, language:ru, language:ko

CLaS-Bench - Cross-Lingual Alignment and Steering Benchmark CLaS-Bench is a lightweight parallel-question benchmark for evaluating language-forcing / language-steering behaviour in large language models (LLMs) across 32 typologically diverse languages. Beyond language steering and alignment evaluation, the dataset can also be used for a variety of multilingual research tasks, including cross-lingual transfer, language identification, representation analysis, multilingual… See the full description on the dataset page: https://huggingface.co/datasets/DGurgurov/CLaS-Bench.

View on Hugging FaceReport an issue

Author / organization	DGurgurov
License (as declared)	cc-by-4.0
Downloads	473 (as reported by Hugging Face)
Likes	1
Created	25 February 2026
Last updated at source	9 October 2026

multilinguallanguage-steeringmechanistic-interpretabilitylanguage-model-evaluationparallel-corpus

Embed this entry

<iframe src="https://silytics.si/embed/item/7718d12b-6a61-44d0-ba75-7c9939e73749" width="420" height="190" style="border:0" loading="lazy" title="Silytics entry"></iframe>

Free, no tracking. More about feeds and embeds.

Popularity, last 3 days

473 now (-464 since 2026-10-16). Counts as reported by the source, snapshotted daily by Silytics.

More from DGurgurov

Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 18 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.