Pith. sign in

REVIEW 2 major objections 4 minor 47 references

A human-LLM collaboration framework builds country-specific Spanish stereotype datasets at low cost, exposing large regional differences in how models encode bias.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · grok-4.5

2026-07-10 15:50 UTC pith:SQQZO5W2

load-bearing objection Useful released multi-country Spanish stereotype dataset plus a practical LLM-generation + human-validation pipeline; completeness of the candidate pool is the real soft spot, but the work is honest and the empirical results hold up. the 2 major comments →

arxiv 2607.07895 v1 pith:SQQZO5W2 submitted 2026-07-08 cs.CL

Scalable and Culturally Specific Stereotype Dataset Construction via Human-LLM Collaboration

Weicheng Ma , John Guerrerio , Soroush Vosoughi This is my paper
classification cs.CL
keywords stereotype datasetshuman-LLM collaborationcultural biasSpanish NLPcross-cultural evaluationattention-head pruningmultilingual fairness
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

English-only stereotype benchmarks leave non-English cultures understudied because full manual collection is expensive. This paper shows that large language models can first propose candidate stereotypes for a target country, after which local annotators simply validate and turn them into test examples. The resulting EspanStereo dataset covers five Spanish-speaking countries and contains both familiar stereotypes and many that never appear in English resources or prior sociological lists. When Spanish-supporting models are probed with these examples, the attention heads that drive stereotyping differ markedly by country, and pruning the right heads reduces the bias with little harm to language modeling. The same pipeline is language-agnostic, so it offers a practical route to multilingual, culture-grounded bias evaluation.

Core claim

Large language models, when prompted with carefully designed injection attacks, already contain enough cultural knowledge to surface high-quality, country-specific stereotypes; human validation then filters them into a reliable test set. Applied to Spanish, this process yields EspanStereo, whose stereotypes overlap little with English datasets and vary substantially across Spain, Mexico, Argentina, Colombia and Nicaragua. Probing and pruning experiments confirm that the models encode these stereotypes in country-dependent patterns of attention heads.

What carries the argument

Human-LLM collaborative annotation: an LLM first generates candidate stereotypes under constrained, multi-viewpoint prompts; in-culture annotators then rate prevalence on a Likert scale and write context/stereotype/anti-stereotype triples. The validated set becomes EspanStereo.

Load-bearing premise

The stereotypes an LLM can be induced to emit form a sufficiently complete and unbiased sample of the stereotypes that actually circulate in each culture, so human validation mainly removes noise rather than missing whole classes of local bias.

What would settle it

Recruit independent native speakers from the same five countries who have never seen the LLM outputs and ask them to list stereotypes freely; if large numbers of frequently mentioned stereotypes are absent from EspanStereo, the retrieval step is incomplete.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • Spanish-supporting models can now be audited and debiased on a per-country basis rather than with translated English lists.
  • The same generate-then-validate pipeline can be run for any language that has at least modest LLM coverage, producing new culture-specific benchmarks at far lower cost than pure manual collection.
  • Attention-head pruning guided by EspanStereo-style data becomes a practical mitigation tool for regional stereotypes.
  • Future multilingual bias suites can be assembled incrementally, country by country, instead of waiting for large-scale sociological surveys.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • If the method works for Spanish, it should also surface under-documented stereotypes in other mid-resource languages that share training data with major LLMs, such as Portuguese or Turkish.
  • Country-level differences in attention-head rankings suggest that a single global debiasing recipe may be suboptimal; fine-grained cultural adapters could be more effective.
  • The high validation rates imply that LLM pre-training already encodes many local stereotypes, so safety filters that simply block stereotype generation may also hide useful cultural knowledge.
  • Repeating the pipeline every few years with newer models could track how quickly emerging social stereotypes enter model weights.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

2 major / 4 minor

Summary. The paper introduces a human-LLM collaborative framework that uses injection-style prompting of LLMs (from six points of view, repeated to saturation) to generate candidate stereotypes, followed by validation and StereoSet-style instantiation by in-culture annotators. It applies the framework to construct EspanStereo, covering race, religion, gender, sexual orientation and age stereotypes for Spain, Mexico, Argentina, Colombia and Nicaragua (538 validated stereotypes, 2 690 triples). Validation rates are high for most country-category cells (Table 7), overlap with StereoSet/CrowS-Pairs is low (Tables 11-12), many stereotypes are absent from existing sociological literature (Table 10), and country-specific cultural grounding is illustrated (Table 6). Shapley-value probing and attention-head pruning on BETO and XLM-R (following Ma et al. 2023b) show country-dependent contribution patterns (Figure 2) and that top-down pruning moves stereotype scores toward 50 while largely preserving language-modeling scores.

Significance. If the results hold, the work supplies both a concrete multi-country Spanish stereotype benchmark and a language-agnostic, lower-cost construction pipeline that can reduce the annotation burden that has limited non-English resources. Explicit strengths include the public MIT-licensed release of EspanStereo, transparent reporting of validation rates, inter-annotator vote ratios (Tables E1-E5), literature and English-dataset overlaps, and faithful reproduction of a published probing protocol that yields the expected top-down versus bottom-up ablation curves. These elements make the contribution immediately usable for culturally grounded evaluation of Spanish-supporting models and for scaling similar resources to other languages.

major comments (2)
  1. [Section 3.1, Table 7, Limitations] The central claim that the framework yields high-quality, country-specific stereotypes largely absent from English resources and literature rests on the untested completeness of the LLM-generated candidate pool (Section 3.1). High validation rates (Table 7, mostly >85 %) and low overlaps (Tables 10-12) establish precision of the emitted set after majority-vote filtering (median Likert ≤ 2 discarded), but supply no recall measure against stereotypes actually circulating in each culture. The Nicaragua race cell (36 % validation) shows that annotators successfully reject invalid immigration stereotypes, yet the Limitations section itself notes weaker coverage of less-prominent or emerging stereotypes. Without a complementary human-elicitation baseline or other completeness check for at least one country, the claims of cultural specificity, low English overlap, and country-dependent encoding
  2. [Section 6, Figure 2] The probing and pruning experiments that support the country-variation claim (Section 6, Figure 2) are conducted only on two encoder models (BETO, XLM-R). While the protocol is correctly followed and the ablation curves behave as expected, the paper repeatedly frames its contribution in terms of LLMs more broadly; the absence of even one modern decoder-only Spanish-supporting model leaves open whether the observed country-dependent attention-head patterns generalize beyond the two tested architectures.
minor comments (4)
  1. [Title page and references] Several author names appear with anomalous spacing (e.g., "V osoughi", "Soroush V osoughi"); these should be corrected throughout.
  2. [Figures 2, F1-F6] Figures F1-F6 and the correlation heatmaps in Figure 2 would benefit from larger fonts and explicit color-bar legends so that positive versus negative Shapley values remain legible in print.
  3. [Section 3.1, Ethics Statement] The term "injection attack" is used for the prompting strategy; a more neutral description (e.g., "adversarial role-play prompting") would better match the ethics discussion and avoid unnecessary security connotations.
  4. [Tables 1-5] Table 1-5 captions refer to "Proportion of Mexican stereotypes shared by other countries" etc.; a short note clarifying that the percentages are computed after validation would remove ambiguity.

Circularity Check

1 steps flagged

No circular reduction in the dataset-construction or evaluation claims; only a non-load-bearing self-citation of the authors' prior probing method.

specific steps
  1. self citation load bearing [Section 6 (Stereotype Examination & Mitigation)]
    "we utilized the methodology proposed by Ma et al. (2023b), which quantifies the contributions of attention heads in transformer-based LLMs using Shapley value-based probing and reduces stereotypes through attention-head pruning."

    The evaluation of country-dependent encoding patterns and the claim that EspanStereo enables mitigation rest on the authors' own prior probing method. The citation is not load-bearing for the dataset-construction claims themselves (those rest on independent human validation), so the circularity is minor and does not force the central results by definition.

full rationale

The paper's core pipeline (LLM candidate generation via injection prompts, independent in-culture human validation on a Likert scale with majority-vote filtering, human instantiation into StereoSet-style triples, and subsequent country-wise comparison against literature and English datasets) is self-contained and does not define any output quantity in terms of a fitted parameter that is later re-presented as a prediction. Validation rates, overlap tables, and cultural-specificity claims rest on external human judgments and external corpora, not on self-referential equations. The sole self-citation (Ma et al. 2023b) supplies the Shapley-value probing/pruning procedure used in Section 6; that procedure is applied to the newly constructed EspanStereo data and is not invoked to justify the dataset's existence, completeness, or cultural fidelity. No uniqueness theorem, ansatz, or renaming of a known empirical pattern is smuggled in. Consequently the derivation chain contains no circular step that reduces a claimed result to its own inputs by construction.

Axiom & Free-Parameter Ledger

4 free parameters · 3 axioms · 0 invented entities

The central claims rest on a small set of methodological choices (Likert threshold, majority vote, number of annotators, injection-prompt design) and on the domain premise that stereotypes frequent enough to appear in LLM training data are recoverable by prompting. No new physical or mathematical entities are postulated; free parameters are the usual experimental knobs of annotation and probing.

free parameters (4)
  • Likert median discard threshold = ≤2
    Stereotypes with median score ≤2 are discarded; the cut-off is chosen by the authors and directly determines the final 538-stereotype inventory.
  • Number of in-culture annotators per country = 5
    Fixed at five for both validation and instantiation; majority vote over this fixed panel decides inclusion.
  • Shapley sampling rate = 256
    256 samples used for attention-head contribution estimates; affects ranking stability.
  • Prediction-head learning rate = 1e-4
    1e-4 used when training the stereotype-detection heads for probing.
axioms (3)
  • domain assumption Stereotypes that are widely held in a culture appear frequently enough in LLM pre-training data to be elicited by suitably crafted prompts.
    Stated in Section 3 as the justification for using LLMs as the source of candidate stereotypes.
  • ad hoc to paper Majority vote of five in-culture annotators on a 5-point Likert scale is a sufficient filter for cultural authenticity.
    Adopted in Section 3.2; the paper acknowledges known limitations of inter-annotator agreement for subjective stereotype tasks but does not replace the majority-vote rule.
  • ad hoc to paper Injection prompts that force the model to list both positive and negative stereotypes (or to role-play unsavory speakers) reliably surface the stereotypes present in the model without introducing large numbers of hallucinations.
    Core of the generation procedure in Section 3.1 and Appendix A; different models require different prompt variants.

pith-pipeline@v1.1.0-grok45 · 28527 in / 2737 out tokens · 33205 ms · 2026-07-10T15:50:21.845595+00:00 · methodology

0 comments
read the original abstract

Research on stereotypes in large language models (LLMs) has largely focused on English-speaking contexts, due to the lack of datasets in other languages and the high cost of manual annotation in underrepresented cultures. To address this gap, we introduce a cost-efficient human-LLM collaborative annotation framework and apply it to construct EspanStereo, a Spanish-language stereotype dataset spanning multiple Spanish-speaking countries across Europe and Latin America. EspanStereo captures both well-documented stereotypes from prior literature and culturally specific biases absent from English-centric resources. Using LLMs to generate candidate stereotypes and in-culture annotators to validate them, we demonstrate the framework's effectiveness in identifying nuanced, region-specific biases. Our evaluation of Spanish-supporting LLMs using EspanStereo reveals significant variation in stereotypical behavior across countries, highlighting the need for more culturally grounded assessments. Beyond Spanish, our framework is adaptable to other languages and regions, offering a scalable path toward multilingual stereotype benchmarks. This work broadens the scope of stereotype analysis in LLMs and lays the groundwork for comprehensive cross-cultural bias evaluation.

Figures

Figures reproduced from arXiv: 2607.07895 by John Guerrerio, Soroush Vosoughi, Weicheng Ma.

Figure 1
Figure 1. Figure 1: An example injection attack against GPT [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Spearman correlation coefficients between the [PITH_FULL_IMAGE:figures/full_fig_p009_2.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

47 extracted references · 47 canonical work pages · 1 internal anchor

  1. [1]

    and Raimundo Schulz, Emma and Saci, Thiziri and Saidi, Sarah and Torroba Marchante, Javier and Xie, Shilin and Zanotto, Sergio E

    Fort, Karen and Alonso Alemany, Laura and Benotti, Luciana and Bezan c on, Julien and Borg, Claudia and Borg, Marthese and Chen, Yongjian and Ducel, Fanny and Dupont, Yoann and Ivetta, Guido and Li, Zhijian and Mieskes, Margot and Naguib, Marco and Qian, Yuyan and Radaelli, Matteo and Schmeisser-Nieto, Wolfgang S. and Raimundo Schulz, Emma and Saci, Thizi...

    work page 2024

  2. [2]

    S ee GULL Multilingual: a Dataset of Geo-Culturally Situated Stereotypes

    Bhutani, Mukul and Robinson, Kevin and Prabhakaran, Vinodkumar and Dave, Shachi and Dev, Sunipa. S ee GULL Multilingual: a Dataset of Geo-Culturally Situated Stereotypes. Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers). 2024. doi:10.18653/v1/2024.acl-short.75

    work page doi:10.18653/v1/2024.acl-short.75 2024

  3. [3]

    On Evaluating and Mitigating Gender Biases in Multilingual Settings

    Vashishtha, Aniket and Ahuja, Kabir and Sitaram, Sunayana. On Evaluating and Mitigating Gender Biases in Multilingual Settings. Findings of the Association for Computational Linguistics: ACL 2023. 2023. doi:10.18653/v1/2023.findings-acl.21

    work page doi:10.18653/v1/2023.findings-acl.21 2023

  4. [4]

    Investigating Bias in Multilingual Language Models: Cross-Lingual Transfer of Debiasing Techniques

    Reusens, Manon and Borchert, Philipp and Mieskes, Margot and De Weerdt, Jochen and Baesens, Bart. Investigating Bias in Multilingual Language Models: Cross-Lingual Transfer of Debiasing Techniques. Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing. 2023. doi:10.18653/v1/2023.emnlp-main.175

    work page doi:10.18653/v1/2023.emnlp-main.175 2023

  5. [5]

    F rench C row S -Pairs: Extending a challenge dataset for measuring social bias in masked language models to a language other than E nglish

    N. F rench C row S -Pairs: Extending a challenge dataset for measuring social bias in masked language models to a language other than E nglish. Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2022. doi:10.18653/v1/2022.acl-long.583

    work page doi:10.18653/v1/2022.acl-long.583 2022

  6. [6]

    Comparing Biases and the Impact of Multilingual Training across Multiple Languages

    Levy, Sharon and John, Neha and Liu, Ling and Vyas, Yogarshi and Ma, Jie and Fujinuma, Yoshinari and Ballesteros, Miguel and Castelli, Vittorio and Roth, Dan. Comparing Biases and the Impact of Multilingual Training across Multiple Languages. Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing. 2023. doi:10.18653/v1/2023...

    work page doi:10.18653/v1/2023.emnlp-main.634 2023

  7. [7]

    All Languages Matter: On the Multilingual Safety of LLM s

    Wang, Wenxuan and Tu, Zhaopeng and Chen, Chang and Yuan, Youliang and Huang, Jen-tse and Jiao, Wenxiang and Lyu, Michael. All Languages Matter: On the Multilingual Safety of LLM s. Findings of the Association for Computational Linguistics: ACL 2024. 2024. doi:10.18653/v1/2024.findings-acl.349

    work page doi:10.18653/v1/2024.findings-acl.349 2024

  8. [8]

    , year =

    Nangia, Nikita and Vania, Clara and Bhalerao, Rasika and Bowman, Samuel R. C row S -Pairs: A Challenge Dataset for Measuring Social Biases in Masked Language Models. Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP). 2020. doi:10.18653/v1/2020.emnlp-main.154

    work page doi:10.18653/v1/2020.emnlp-main.154 2020

  9. [9]

    Deciphering Stereotypes in Pre-Trained Language Models

    Ma, Weicheng and Scheible, Henry and Wang, Brian and Veeramachaneni, Goutham and Chowdhary, Pratim and Sun, Alan and Koulogeorge, Andrew and Wang, Lili and Yang, Diyi and Vosoughi, Soroush. Deciphering Stereotypes in Pre-Trained Language Models. Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing. 2023. doi:10.18653/v1/2...

    work page doi:10.18653/v1/2023.emnlp-main.697 2023

  10. [10]

    Unsupervised Cross-lingual Representation Learning at Scale

    Conneau, Alexis and Khandelwal, Kartikay and Goyal, Naman and Chaudhary, Vishrav and Wenzek, Guillaume and Guzm \'a n, Francisco and Grave, Edouard and Ott, Myle and Zettlemoyer, Luke and Stoyanov, Veselin. Unsupervised Cross-lingual Representation Learning at Scale. Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics. ...

    work page doi:10.18653/v1/2020.acl-main.747 2020

  11. [11]

    Praeger guide to the psychology of gender , pages=

    A Latin American perspective on the study of gender , author=. Praeger guide to the psychology of gender , pages=. 2004 , publisher=

    work page 2004

  12. [12]

    The Wiley Encyclopedia of Personality and Individual Differences: Clinical, Applied, and Cross-Cultural Research , pages=

    Marianismo , author=. The Wiley Encyclopedia of Personality and Individual Differences: Clinical, Applied, and Cross-Cultural Research , pages=. 2020 , publisher=

    work page 2020

  13. [13]

    Journal of Multicultural Social Work , volume=

    Machismo, fatherhood and the Latino family: Understanding the concept , author=. Journal of Multicultural Social Work , volume=. 1997 , publisher=

    work page 1997

  14. [14]

    American Journal of Sociology , volume=

    Understanding Latin American beliefs about racial inequality , author=. American Journal of Sociology , volume=. 2013 , publisher=

    work page 2013

  15. [15]

    2014 , publisher=

    Race and ethnicity in Latin American history , author=. 2014 , publisher=

    work page 2014

  16. [16]

    Proceedings of the 59th

    Nadeem, Moin and Bethke, Anna and Reddy, Siva. S tereo S et: Measuring stereotypical bias in pretrained language models. Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers). 2021. doi:10.18653/v1/2021.acl-long.416

    work page doi:10.18653/v1/2021.acl-long.416 2021

  17. [17]

    Gender Bias in Coreference Resolution: Evaluation and Debiasing Methods

    Zhao, Jieyu and Wang, Tianlu and Yatskar, Mark and Ordonez, Vicente and Chang, Kai-Wei. Gender Bias in Coreference Resolution: Evaluation and Debiasing Methods. Proceedings of the 2018 Conference of the North A merican Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 2 (Short Papers). 2018. doi:10.18653/v1/N18-2003

    work page doi:10.18653/v1/n18-2003 2018

  18. [18]

    W ino Q ueer: A Community-in-the-Loop Benchmark for Anti- LGBTQ + Bias in Large Language Models

    Felkner, Virginia and Chang, Ho-Chun Herbert and Jang, Eugene and May, Jonathan. W ino Q ueer: A Community-in-the-Loop Benchmark for Anti- LGBTQ + Bias in Large Language Models. Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2023. doi:10.18653/v1/2023.acl-long.507

    work page doi:10.18653/v1/2023.acl-long.507 2023

  19. [19]

    PML4DC at ICLR 2020 , year=

    Spanish Pre-Trained BERT Model and Evaluation Data , author=. PML4DC at ICLR 2020 , year=

    work page 2020

  20. [20]

    Joint European Conference on Machine Learning and Knowledge Discovery in Databases , pages=

    How Different is Stereotypical Bias Across Languages? , author=. Joint European Conference on Machine Learning and Knowledge Discovery in Databases , pages=. 2023 , organization=

    work page 2023

  21. [21]

    A methodology to characterize bias and harmful stereotypes in natural language processing in Latin America

    A methodology to characterize bias and harmful stereotypes in natural language processing in Latin America , author=. arXiv preprint arXiv:2207.06591 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  22. [22]

    First Conference on Language Modeling , year=

    Vera Neplenbroek and Arianna Bisazza and Raquel Fern. First Conference on Language Modeling , year=

    work page

  23. [23]

    and Dave, Shachi and Prabhakaran, Vinodkumar and Dev, Sunipa

    Jha, Akshita and Davani, Aida and Reddy, Chandan K. and Dave, Shachi and Prabhakaran, Vinodkumar and Dev, Sunipa. S ee GULL : A Stereotype Benchmark with Broad Geo-Cultural Coverage Leveraging Generative Models. Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2023. doi:10.18653/v1/2023.acl-long.548

    work page doi:10.18653/v1/2023.acl-long.548 2023

  24. [24]

    Journal of Artificial Intelligence Research , volume=

    Learning from disagreement: A survey , author=. Journal of Artificial Intelligence Research , volume=

    work page

  25. [25]

    Intersectional Stereotypes in Large Language Models: Dataset and Analysis

    Ma, Weicheng and Chiang, Brian and Wu, Tong and Wang, Lili and Vosoughi, Soroush. Intersectional Stereotypes in Large Language Models: Dataset and Analysis. Findings of the Association for Computational Linguistics: EMNLP 2023. 2023. doi:10.18653/v1/2023.findings-emnlp.575

    work page doi:10.18653/v1/2023.findings-emnlp.575 2023

  26. [26]

    Educational Research Bulletin , pages=

    A technique for the measurement of attitudes in the social studies , author=. Educational Research Bulletin , pages=. 1936 , publisher=

    work page 1936

  27. [27]

    Perspectives on psychological science , volume=

    Prejudices in cultural contexts: Shared stereotypes (gender, age) versus variable stereotypes (race, ethnicity, religion) , author=. Perspectives on psychological science , volume=. 2017 , publisher=

    work page 2017

  28. [28]

    Archetti, Eduardo and Olivari, Nicol. The. The Buenos Aires Reader: History, Culture, Politics , pages =. 2024 , doi =

    work page 2024

  29. [29]

    Environment and Urbanization , volume=

    Local government and decentralization in Colombia , author=. Environment and Urbanization , volume=. 1991 , publisher=

    work page 1991

  30. [30]

    The rights of the Way

    Ulloa, Astrid , journal=. The rights of the Way. 2020 , publisher=

    work page 2020

  31. [31]

    El Indio

    Manifest destiny, the frontier, and “El Indio” in Argentina’s conquista del desierto , author=. Journal of Global South Studies , volume=. 2019 , publisher=

    work page 2019

  32. [32]

    Journal of Latin American Geography , pages=

    Indigenous territoriality at the end of the social property era in Mexico , author=. Journal of Latin American Geography , pages=. 2010 , publisher=

    work page 2010

  33. [33]

    Cardellino, Cristian , title =

    work page

  34. [34]

    Social Sciences , volume=

    The persistence of racial constructs in Spain: Bringing race and colorblindness into the debate on interculturalism , author=. Social Sciences , volume=. 2022 , publisher=

    work page 2022

  35. [35]

    The spanish journal of Psychology , volume=

    Evolution of gender stereotypes in Spain: Traits and roles , author=. The spanish journal of Psychology , volume=. 2008 , publisher=

    work page 2008

  36. [36]

    Geriatrics , volume=

    The Spanish intergenerational study: Beliefs, stereotypes, and metacognition about older people and grandparents to tackle ageism , author=. Geriatrics , volume=. 2021 , publisher=

    work page 2021

  37. [37]

    Frontiers in Psychology , volume=

    Correlate attitudes toward LGBT and sexism in Spanish psychology students , author=. Frontiers in Psychology , volume=. 2020 , publisher=

    work page 2020

  38. [38]

    Social Compass , volume=

    The silent inquisition: Islamophobic microaggressions and Spanish Moroccan identity negotiations in contemporary Madrid , author=. Social Compass , volume=. 2022 , publisher=

    work page 2022

  39. [39]

    Ageing International , volume =

    Torres, Carmen , title =. Ageing International , volume =

    work page

  40. [40]

    2010 , publisher=

    The politics of sexuality in Latin America: A reader on lesbian, gay, bisexual, and transgender rights , author=. 2010 , publisher=

    work page 2010

  41. [41]

    Cold War History , pages=

    Contra solidarity: revolution and counterrevolution in the united states and nicaragua , author=. Cold War History , pages=. 2024 , publisher=

    work page 2024

  42. [42]

    David , booktitle =

    Parrish, Charles and Tyler, B. David , booktitle =. Supercl. 2021 , publisher=

    work page 2021

  43. [43]

    Latin American Research Review , volume=

    Religion and society in Latin America: Ambivalence and advances , author=. Latin American Research Review , volume=. 1996 , publisher=

    work page 1996

  44. [44]

    Strengthening Romani Voices in Colombia: Reflections on a Participatory Approach , journal =

    Marsh, Hazel and Acu. Strengthening Romani Voices in Colombia: Reflections on a Participatory Approach , journal =

    work page

  45. [45]

    Revealing the Dark Secrets of BERT

    Kovaleva, Olga and Romanov, Alexey and Rogers, Anna and Rumshisky, Anna. Revealing the Dark Secrets of BERT. Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP). 2019. doi:10.18653/v1/D19-1445

    work page doi:10.18653/v1/d19-1445 2019

  46. [46]

    SHADES : Towards a Multilingual Assessment of Stereotypes in Large Language Models

    Mitchell, Margaret and Attanasio, Giuseppe and Baldini, Ioana and Clinciu, Miruna and Clive, Jordan and Delobelle, Pieter and Dey, Manan and Hamilton, Sil and Dill, Timm and Doughman, Jad and Dutt, Ritam and Ghosh, Avijit and Forde, Jessica Zosa and Holtermann, Carolin and Kaffee, Lucie-Aim \'e e and Laud, Tanmay and Lauscher, Anne and Lopez-Davila, Rober...

    work page doi:10.18653/v1/2025.naacl-long.600 2025

  47. [47]

    Building Stereotype Repositories with Complementary Approaches for Scale and Depth

    Dev, Sunipa and Jha, Akshita and Goyal, Jaya and Tewari, Dinesh and Dave, Shachi and Prabhakaran, Vinodkumar. Building Stereotype Repositories with Complementary Approaches for Scale and Depth. Proceedings of the First Workshop on Cross-Cultural Considerations in NLP (C3NLP). 2023. doi:10.18653/v1/2023.c3nlp-1.9

    work page doi:10.18653/v1/2023.c3nlp-1.9 2023