Pith. sign in

REVIEW 6 cited by

MBBQ: A Dataset for Cross-Lingual Comparison of Stereotypes in Generative LLMs

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2406.07243 v3 pith:AB2TCK47 submitted 2024-06-11 cs.CL

classification cs.CL
keywords biasmbbqdatasetlanguagesllmsmodelsstereotypesenglish
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Generative large language models (LLMs) have been shown to exhibit harmful biases and stereotypes. While safety fine-tuning typically takes place in English, if at all, these models are being used by speakers of many different languages. There is existing evidence that the performance of these models is inconsistent across languages and that they discriminate based on demographic factors of the user. Motivated by this, we investigate whether the social stereotypes exhibited by LLMs differ as a function of the language used to prompt them, while controlling for cultural differences and task accuracy. To this end, we present MBBQ (Multilingual Bias Benchmark for Question-answering), a carefully curated version of the English BBQ dataset extended to Dutch, Spanish, and Turkish, which measures stereotypes commonly held across these languages. We further complement MBBQ with a parallel control dataset to measure task performance on the question-answering task independently of bias. Our results based on several open-source and proprietary LLMs confirm that some non-English languages suffer from bias more than English, even when controlling for cultural shifts. Moreover, we observe significant cross-lingual differences in bias behaviour for all except the most accurate models. With the release of MBBQ, we hope to encourage further research on bias in multilingual settings. The dataset and code are available at https://github.com/Veranep/MBBQ.

Discussion (0). Sign in to comment.

Forward citations

Cited by 6 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. StereoTales: A Multilingual Framework for Open-Ended Stereotype Discovery in LLMs

    cs.CY 2026-05 unverdicted novelty 7.0 of 10

    StereoTales shows that LLMs produce harmful, culturally adapted stereotypes in open-ended multilingual stories, with patterns consistent across providers and aligned human-LLM harm judgments.

  2. StereoTales: A Multilingual Framework for Open-Ended Stereotype Discovery in LLMs

    cs.CY 2026-05 accept novelty 7.0 of 10

    StereoTales shows that all tested LLMs emit harmful stereotypes in open-ended stories, with associations adapting to prompt language and targeting locally salient groups rather than transferring uniformly across languages.

  3. Cross-Lingual Bias in Large Language Models: A Comparative Analysis of English and Swahili

    cs.CL 2026-08 conditional novelty 6.0 of 10

    Bias in GPT-5.2 and Gemini 2.5 Flash changes rather than transfers between English and Swahili, with GPT-5.2 refusal behavior appearing only in English.

  4. QuantiBias: Benchmarking Quantization-Induced Bias in LLMs

    cs.CL 2026-07 conditional novelty 6.0 of 10

    Quantization leaves refusal and multiple-choice bias checks flat while open-ended stereotype endorsement remains high (~24–27% under an independent judge), a gap standard safety evaluations miss.

  5. BharatBBQ: A Multilingual Bias Benchmark for Question Answering in the Indian Context

    cs.CL 2025-08 conditional novelty 6.0 of 10

    BharatBBQ measures social bias in question-answering models across eight languages and finds that Indian-language examples often elicit more stereotyped answers than English ones.

  6. Preserving Fairness and Safety in Quantized LLMs Through Critical Weight Protection

    cs.CL 2026-01 conditional novelty 5.0 of 10

    Quantization tends to degrade LLM fairness and safety—more in non-English tasks—and preserving top sensitivity-ranked weights in FP16 mostly mitigates the loss.

Pith tools