Pith. sign in

REVIEW 3 major objections 4 minor 1 cited by

Small Changes, Large Consequences: Analyzing the Allocational Fairness of LLMs in Hiring Contexts

T0 review · 3 major / 4 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read An LLM hiring pipeline that retrieves and summarizes resumes is allocationally unfair, with the retrieval stage as the main source of bias.

desk verdict Solid benchmark and new metric, but the retrieval-bias claim rests on untested asymmetries; deserves review with revisions. read the letter →

arxiv 2501.04316 v2 pith:A7UJRSX2 submitted 2025-01-08 cs.CL

Preethi Seshadri , Hongyu Chen , Sameer Singh , Seraphina Goldfarb-Tarrant This is my paper · ORCID
classification cs.CL
keywords allocationalfairnessLLMhiringbiasresumesummarizationretrievaldemographicperturbationinvarianceviolationsexclusionmetricsyntheticbenchmark
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper asks whether an LLM-based hiring pipeline—resume retrieval followed by resume summarization—treats otherwise identical candidates differently when their demographic signals change. Using a new synthetic resume benchmark with controlled name and extracurricular perturbations, it finds that race changes alter generated summaries up to 20% of the time (gender only 3%), and that retrieval models exclude demographically perturbed resumes from top-5 lists nearly half the time on average, up to 74% in specific settings. The authors argue the harm is real: candidates who differ only by name can be filtered out or described less favorably, and the retrieval stage is the larger source of bias. A striking additional finding is that non-demographic changes—typos and spacing—produce comparable ranking shifts, suggesting much of the measured unfairness is general brittleness rather than targeted demographic bias.

What carries the argument

The central object is the paired perturbation design: each resume has demographically perturbed twins that keep content fixed, so any output difference is attributable to demographic signal. For summarization, the paper counts invariance violations—the fraction of paired t-tests (with multiple-comparison correction) that reject equality across five automated text measures (reading ease, reading time, polarity, subjectivity, regard). For retrieval, it proposes the exclusion metric—the fraction of top-n resumes that fall out of the top-n set after perturbation—alongside the non-uniformity chi-square test of demographic composition in top-x% results. These metrics turn the abstract question 'is the system fair?' into concrete per-stage measurements that can be compared across models and perturbation types.

What would settle it

Have HR professionals shortlist candidates from paired summaries of identical resumes that differ only by demographic perturbation; if summaries with statistically different reading ease, polarity, or regard do not produce different shortlisting rates, the summarization-stage bias does not translate into allocational harm. For the retrieval stage, compare exclusion rates under demographic name swaps against exclusion under frequency-matched but demographically uninformative name swaps; if the control shows equally high exclusion, the observed 'bias' is general brittleness rather than demographic discrimination.

Watch

Extended reading notes

Core claim

The paper claims that an LLM hiring pipeline with automated resume retrieval and summarization is not allocationally fair, and that the retrieval stage is the primary driver. On identical resumes differing only by gender- or race-typed names (sometimes plus demographically tailored extracurriculars), generated summaries show statistically meaningful differences far more often for race than for gender—up to 20% versus about 3%—and retrieval models select resumes non-uniformly across demographic groups and drop perturbed resumes out of the top-n pool at high rates (averaging 45.75% for top-5, up to 74% in specific settings). The paper further claims that non-demographic perturbations (typos, spacing, within-group name changes) cause similar or even higher exclusion, so much of the apparent demographic bias is better described as general model brittleness. In the worst composite case the authors report roughly 70% of candidates filtered at screening and about 20% of remaining candidates depicted less favorably after racial indicators change.

Load-bearing premise

The summarization-stage finding rests on the assumption that the five automated text measures (reading ease, reading time, polarity, subjectivity, regard) capture differences that would actually change a recruiter's assessment; the paper validates that HR professionals have preferences when summaries differ on other characteristics, but never shows that demographic-induced differences in these measures alter real hiring decisions.

Editorial extensions

If this is right

  • In a standard retrieval-then-summarize hiring pipeline, the retrieval stage is the main source of allocational harm; fixing only the summarizer would leave most of the bias in place.
  • Because swapping only a candidate's name can drop a resume from the top-5 set nearly half the time on average, screening systems that filter to small candidate pools can silently exclude qualified applicants from underrepresented groups.
  • Race perturbations change generated summaries more often than gender perturbations do, so summary-based screening can portray otherwise identical candidates differently along racial lines.
  • The high sensitivity to typos and spacing means some apparent demographic bias is actually general model brittleness; audits should control for non-demographic perturbations before attributing exclusion to demographics.
  • Adding demographically tailored extracurricular information increases exclusion asymmetrically, especially for White-to-Black name changes, so real resumes with multiple demographic signals may produce worse disparities than name-only tests suggest.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Because the paper finds comparable exclusion for within-group name changes and formatting edits, the exclusion metric may be measuring embedding-model instability as much as demographic bias; a fair-system audit should always pair exclusion with a non-demographic control before attributing disparity to protected attributes.
  • The 20%-versus-3% race/gender gap in summarization suggests race-linked content cues are more strongly encoded in resume text than gender cues; testing intersectional race-gender perturbations would likely show effects that are not additive.
  • Directional asymmetries (e.g., larger White-to-Black than Black-to-White exclusion) imply that fairness interventions need to be direction-aware; a symmetric swap design, as used here, is the right auditing default but should be extended to more groups and to non-name demographic signals such as age and nationality.
  • The dependence of exclusion magnitude on resume formatting (structured generated resumes were nearly twice as sensitive to spacing changes as web-scraped ones) suggests that standardizing resume formats before retrieval could reduce both apparent and real disparities.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. The paper studies allocational fairness of LLM-based hiring pipelines using two tasks: resume summarization and candidate retrieval/ranking. The authors construct a synthetic resume dataset from real resumes, adding four demographic name groups (Black/White male/female) and, in some conditions, demographically tailored extracurricular content. They define invariance violations for summarization (the percentage of paired t-tests rejecting equality across a battery of text-proxy measures) and exclusion for retrieval (the percentage of top-n candidates dropping out after demographic perturbation), together with the non-uniformity metric of Wilson and Caliskan (2024). They evaluate six summarization models and four embedding models. Results: race-perturbed summaries show more invariance violations than gender; retrieval models show high exclusion for both gender and race perturbations, with directional asymmetries; non-demographic perturbations (within-group names, typos, spacing) cause comparable sensitivity. The paper concludes that LLM hiring pipelines are brittle and can exhibit considerable bias, primarily in the retrieval stage.

Significance. If the findings hold, this is a valuable, ecologically grounded contribution: it moves fairness evaluation from classification tasks to generative and retrieval settings, introduces reusable metrics (invariance violations, exclusion) and a public benchmark, and includes a human preference validation for the summarization proxies. The paper also honestly reports that non-demographic perturbations produce similar exclusion magnitudes, which is an interesting robustness finding. The benchmark and code release will be useful to the community. The metrics are defined without fitted free parameters (aside from the standard alpha significance level), and the central empirical results are reproducible in principle from the released artifacts. Main weaknesses: the directional asymmetry claims lack uncertainty quantification, and the name perturbation design does not control for tokenization length or lexical form, both of which could affect the central retrieval conclusion. The summarization proxy measures are validated only on artificially manipulated dimensions, not on demographic perturbations, so the allocational interpretation of the summarization results is provisional.

major comments (3)
  1. [§4.2.2, Figure 5] The paper's central claim that retrieval models exhibit "considerable bias, primarily stemming from the retrieval stage" rests on the directional asymmetry of exclusion (e.g., mistral-embed M→F 63.28% vs. F→M 27.93%; W→B vs. B→W trends). These values are presented as point estimates aggregated over job postings without confidence intervals or significance tests. Because Figure 7 shows that non-demographic perturbations (within-group name changes, spacing, typos) produce exclusion rates of comparable magnitude, the demographic asymmetry is the only evidence distinguishing demographic bias from generic brittleness. Without a statistical test of the directional differences (e.g., bootstrapping over job postings or a paired permutation test), the observed asymmetries may be within sampling noise, and the "considerable bias" conclusion is not yet established.
  2. [§3.3 and §3.2] Name perturbations are matched on frequency (via Pile bins) but not on tokenization length or lexical surface form. The authors cite An and Rudinger (2023) in the related work, which shows that tokenization length alone can produce first-name bias in model outputs. Since the Yin et al. (2024) name sets differ systematically in length and spelling patterns (e.g., "Latanya" vs. "Emily"), the reported directional effects W→B > B→W (Figure 5) and the sensitivity to within-group name changes (Figure 7) could be confounded by tokenization artifacts. The manuscript should either match names across groups on tokenization length (in addition to frequency) or include token length as a covariate in a per-resume logistic regression predicting exclusion from perturbation direction. This is necessary to attribute the retrieval asymmetry to demographic group membership.
  3. [§2.1 and §A.9] The validation of the summarization proxy measures does not bridge the gap from text-property differences to allocational harm. The human preference study shows that HR professionals have preferences when summaries differ along deliberately manipulated dimensions (quantification, focus, individual impact) and that the five automated measures detect those manipulated differences. However, the invariance violations reported in Figure 2 are computed on demographic perturbations, not on the manipulated dimensions; there is no evidence that the demographic-induced changes in reading ease, polarity, regard, etc. correspond to changes that would affect an HR decision. Consequently, the Discussion's claim that "20% of remaining candidates being depicted in less preferable ways" (Section 5) overstates what the data show. The summarization results should be reported as differences in proxy measures, with the downstream allocational interpretation clearly flagged as an assumption awaiting validation.
minor comments (4)
  1. [§2.2, Eq. (2)] The displayed formula for exclusion is malformed; it should read exclusion_n(p) = |{d ∈ D_n(p) : rank_p(d′) > n}| / |D_n(p)|.
  2. [Figure 2] Figure 2 is visually cluttered and difficult to compare across models; consider supplementing it with a table of numeric invariance-violation values per model and perturbation type.
  3. [§3.1 and §3.4] The limitations section should explicitly discuss the fact that Command-R generated the seed resumes and is also evaluated as a summarization model; this coupling could introduce artifacts for that model and should be acknowledged as a potential confound in the summarization results.
  4. [§3.1] The paper does not justify why 11 professions/fields were selected for job posts from the 22/24 available; please state the selection criterion or note that the choice was based on job-posting availability.

Circularity Check

0 steps flagged · score 1.0 of 10

No circular derivation: the fairness metrics are direct operational measurements, and the paper's conclusions do not reduce to fitted inputs or self-citation chains.

full rationale

The paper's claims are empirical measurements, not derived predictions. Summarization 'invariance violations' are defined as the proportion of paired t-tests rejecting equality between original and perturbed summaries over five surface proxies; the reported RQ1 result is a direct readout of this operationalization, not a prediction fitted from the data. The retrieval 'exclusion' metric is likewise a direct counting of top-n rank drops after perturbation; the central finding that demographic and non-demographic perturbations produce comparable exclusion is an observed comparison, not a quantity forced by the metric's definition. No parameter is fitted to a subset of data and then 'predicted' on a closely related quantity. The proxies were validated by an external HR-preference study, which evaluates whether the measures detect human-meaningful differences; even if that validation is incomplete, it is not circular. The paper's minor self-citations (Ribeiro et al. 2020 for invariance framing; Elazar et al. 2023 for name-frequency lookups; Goldfarb-Tarrant et al. 2021 for ecological validity) are not load-bearing: none supplies a result that the present conclusions are equivalent to. The use of Command-R to generate synthetic resumes and also as a summarization model is a potential data-generation confound, but it does not constitute definitional circularity, and the fairness comparisons are within-model across perturbed versions. No 'uniqueness theorem' or equivalent self-citation chain forces any choice. The main epistemic weaknesses—no significance testing on directional asymmetries and the incomplete mapping from proxy differences to recruiter decisions—are correctness/validity concerns, not circularity.

Assumptions & free parameters 1 free parameters · 4 assumptions · 0 invented entities

The paper introduces a new metric (exclusion) and a new dataset, but no new physical or conceptual entities. The main assumptions are about the validity of proxies for demographic signals and hiring relevance. The free parameter alpha is a standard significance threshold, not a fitted value.

free parameters (1)
  • alpha = 0.05
    Significance threshold for paired t-tests in the summarization invariance metric; chosen by convention, not fitted, but directly controls the reported violation rate.
assumptions (4)
  • domain assumption Names are a valid proxy for gender and race in resumes.
    Used throughout Section 3.2 to create demographic groups; the paper acknowledges this is standard practice but notes limitations in the Limitations section.
  • domain assumption Demographically perturbing a resume should not change its relevance to a job posting.
    Stated in Section 2.2 as the basis for the exclusion metric; the paper uses within-group name changes and non-name perturbations as baselines to test this.
  • domain assumption The automated proxy measures (reading ease, polarity, etc.) are valid indicators of meaningful differences in resume summaries for hiring.
    Used in Section 2.1 to define invariance violations; only partially validated by the human study in Appendix A.9.
  • domain assumption Cosine similarity is an appropriate relevance measure for resume retrieval.
    Used for ranking in Section 2.2, following prior work; embedding models are not purpose-built for resume matching.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Small Changes, Large Consequences: Analyzing the Allocational Fairness of LLMs in Hiring Contexts." pith.science (2026). https://pith.science/paper/A7UJRSX2

@misc{pith2026250104316,
  author       = {Pith},
  title        = {Pith review of: Small Changes, Large Consequences: Analyzing the Allocational Fairness of LLMs in Hiring Contexts},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/A7UJRSX2}},
  note         = {Machine review of arXiv:2501.04316}
}
read the original abstract

Large language models (LLMs) are increasingly being deployed in high-stakes applications like hiring, yet their potential for unfair decision-making remains understudied in generative and retrieval settings. In this work, we examine the allocational fairness of LLM-based hiring systems through two tasks that reflect actual HR usage: resume summarization and applicant ranking. By constructing a synthetic resume dataset with controlled perturbations and curating job postings, we investigate whether model behavior differs across demographic groups. Our findings reveal that generated summaries exhibit meaningful differences more frequently for race than for gender perturbations. Models also display non-uniform retrieval selection patterns across demographic groups and exhibit high ranking sensitivity to both gender and race perturbations. Surprisingly, retrieval models can show comparable sensitivity to both demographic and non-demographic changes, suggesting that fairness issues may stem from broader model brittleness. Overall, our results indicate that LLM-based hiring systems, especially in the retrieval stage, can exhibit notable biases that lead to discriminatory outcomes in real-world contexts.

Figures

Figures reproduced from arXiv: 2501.04316 by the authors.

Figure 1
Figure 1. We investigate the fairness of an LLM hiring [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Summarization Results: Invariance violations for generated summaries, separated by completion model and perturbation type. Results are shown across 5 runs. Left 3 models are considered "smaller" models, right 3 models are considered "larger" models. 2020): OpenAI’s text-embedding-3-small and text￾embedding-3-large, Cohere’s embed-english-v3.0, and Mistral’s mistral-embed. 4 Results In this section, we evaluate the u… view at source ↗
Figure 3
Figure 3. Non-uniformity metric for top-5 and top-10% of retrieved resumes. Separated (sep) measures non [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (11 more)
Figure 4
Figure 4. Figure 4: Exclusion metric for retrieval after performing gender and race name perturbations for the top-5, top-10, and top-100 retrieved resumes. Lower values indicate models are less sensitive to demographic perturbations. the time, compared to 5.19% for White females. Reasons…
Figure 5
Figure 5. Figure 5: Directional differences in exclusion metric for retrieval after applying name perturbations (i.e., separating based on perturbation direction). M→F perturbs male to female names and F→M perturbs female to male names, while W→B perturbs White to Black names and B→W pert…
Figure 6
Figure 6. Figure 6: Deltas (differences) in exclusion metric for retrieval after performing demographic perturbations with names + extracurricular information vs. names only. As expected, adding extracurricular information increases sensitivity to perturbations. generated resumes with n =…
Figure 7
Figure 7. Figure 7: Exclusion metric for retrieval after performing non-demographic perturbations (i.e., within group name changes - left, and modifying spacing and adding typos - right). both spacing and typos, but to a lesser extent than names. As shown in Figure 7b, most models demonst…
Figure 8
Figure 8. Figure 8: Human Annotation Results for 3 characteristics (quantification, focus, and individual impact). [PITH_FULL_IMAGE:figures/full_fig_p016_8.png]
Figure 9
Figure 9. Figure 9: Summarization Results: Invariance violations for generated summaries, separated by completion model and perturbation type. Results are shown across 5 runs. Left 3 models are considered "smaller" models, right 3 models are considered "larger" models. Bonferroni correcti…
Figure 10
Figure 10. Figure 10: Non-uniformity metric for top-5 and top-10% of retrieved Kaggle resumes. Separated (sep) measures the [PITH_FULL_IMAGE:figures/full_fig_p018_10.png]
Figure 11
Figure 11. Figure 11: Directional differences in exclusion metric for retrieval (generated resumes) after applying name perturbations (i.e., separating based on perturbation direction). M→F perturbs male names to female names and F→M perturbs female names to male names, while W→B perturbs …
Figure 12
Figure 12. Figure 12: Exclusion metric for retrieval after performing non-demographic perturbations on generated resumes (i.e., within group name changes - top, and modifying spacing and adding typos - bottom). (a) Gender, n = 5 (b) Gender, n = 10 (c) Gender, n = 100 (d) Race, n = 5 (e) Ra…
Figure 13
Figure 13. Figure 13: Directional differences in exclusion metric for retrieval (Kaggle resumes) after applying name perturbations (i.e., separating based on perturbation direction). M→F perturbs male names to female names and F→M perturbs female names to male names, while W→B perturbs Whi…
Figure 14
Figure 14. Figure 14: Exclusion metric for retrieval after performing non-demographic perturbations on Kaggle resumes (i.e., within group name changes - top, and modifying spacing and adding typos - bottom). 20 [PITH_FULL_IMAGE:figures/full_fig_p020_14.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

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

  1. Breaking Down Bias: On The Limits of Generalizable Pruning Strategies

    cs.CL 2025-02 conditional novelty 6.0 of 10

    Pruning-based bias removal in Llama-3-8B reduces racial bias mainly in the context used to choose what to prune, and transfers poorly across contexts.

Reference graph

Works this paper leans on

55 extracted references · 20 canonical work pages · cited by 1 Pith paper

  1. [1]

    online" 'onlinestring :=

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint eprinttype howpublished institution journal key month note number organization pages publisher school series title type volume year doi pubmed url lastchecked label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block STRING...

  2. [2]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...

  3. [3]

    Abubakar Abid, Maheen Farooqi, and James Zou. 2021. https://api.semanticscholar.org/CorpusID:236384212 Large language models associate muslims with violence . Nature Machine Intelligence, 3:461 -- 463

    work page 2021

  4. [4]

    Haozhe An, Christabel Acquaye, Colin Wang, Zongxia Li, and Rachel Rudinger. 2024. https://doi.org/10.18653/v1/2024.acl-short.37 Do large language models discriminate in hiring decisions on the basis of race, ethnicity, and gender? In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), pages 386...

    doi:10.18653/v1/2024.acl-short.37 2024

  5. [5]

    Haozhe An and Rachel Rudinger. 2023. https://doi.org/10.18653/v1/2023.acl-short.34 Nichelle and nancy: The influence of demographic attributes and tokenization length on first name biases . In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), pages 388--401, Toronto, Canada. Association for C...

    doi:10.18653/v1/2023.acl-short.34 2023

  6. [6]

    Solon Barocas, Kate Crawford, Aaron Shapiro, and Hanna Wallach. 2017. The problem with bias: Allocative versus representational harms in machine learning. In 9th Annual conference of the special interest group for computing, information and society, page 1. New York, NY

    work page 2017

  7. [7]

    Yoav Benjamini and Yosef Hochberg. 1995. Controlling the false discovery rate: a practical and powerful approach to multiple testing. Journal of the Royal statistical society: series B (Methodological), 57(1):289--300

    work page 1995

  8. [8]

    Snehaan Bhawal. 2021. Resume dataset. https://www.kaggle.com/datasets/snehaanbhawal/resume-dataset

    work page 2021

Show all 55 references
  1. [9]

    J Martin Bland and Douglas G Altman. 1995. Multiple significance tests: the bonferroni method. Bmj, 310(6973):170

    1995

  2. [10]

    Su Lin Blodgett, Solon Barocas, Hal Daum \'e III, and Hanna Wallach. 2020. https://doi.org/10.18653/v1/2020.acl-main.485 Language (technology) is power: A critical survey of `` bias '' in NLP . In Proceedings of the 58th Annual Meeting of the Association for Computational Ling...

    2020 doi

  3. [11]

    Su Lin Blodgett, Gilsinia Lopez, Alexandra Olteanu, Robert Sim, and Hanna Wallach. 2021. https://doi.org/10.18653/v1/2021.acl-long.81 Stereotyping N orwegian salmon: An inventory of pitfalls in fairness benchmark datasets . In Proceedings of the 59th Annual Meeting of the Asso...

    2021 doi

  4. [12]

    Boston Consulting Group . 2025. https://www.bcg.com/publications/2025/ai-changing-recruitment How ai is changing recruitment

    2025

  5. [13]

    Yang Trista Cao, Yada Pruksachatkun, Kai-Wei Chang, Rahul Gupta, Varun Kumar, Jwala Dhamala, and Aram Galstyan. 2022. https://doi.org/10.18653/v1/2022.acl-short.62 On the intrinsic and extrinsic fairness evaluation metrics for contextualized language representations . In Proce...

    2022 doi

  6. [14]

    Myra Cheng, Esin Durmus, and Dan Jurafsky. 2023. https://doi.org/10.18653/v1/2023.acl-long.84 Marked personas: Using natural language prompts to measure stereotypes in language models . In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics ...

    2023 doi

  7. [15]

    Cohere . 2024. Command-r. https://docs.cohere.com/v2/docs/command-r

    2024

  8. [16]

    Jeffrey Dastin. 2018. https://www.reuters.com/article/world/insight-amazon-scraps-secret-ai-recruiting-tool-that-showed-bias-against-women-idUSKCN1MK0AG Insight - Amazon scraps secret AI recruiting tool that showed bias against women . Reuters

    2018

  9. [17]

    Cynthia Dwork, Moritz Hardt, Toniann Pitassi, Omer Reingold, and Richard Zemel. 2012. https://doi.org/10.1145/2090236.2090255 Fairness through awareness . In Proceedings of the 3rd Innovations in Theoretical Computer Science Conference, ITCS '12, page 214–226, New York, NY, US...

    2012

  10. [18]

    Smith, and Jesse Dodge

    Yanai Elazar, Akshita Bhagia, Ian Helgi Magnusson, Abhilasha Ravichander, Dustin Schwenk, Alane Suhr, Evan Pete Walsh, Dirk Groeneveld, Luca Soldaini, Sameer Singh, Hanna Hajishirzi, Noah A. Smith, and Jesse Dodge. 2023. What's in my big data? In The Twelfth International Conf...

    2023

  11. [19]

    Kawin Ethayarajh, David Duvenaud, and Graeme Hirst. 2019. https://doi.org/10.18653/v1/P19-1166 Understanding undesirable word embedding associations . In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 1696--1705, Florence, Italy....

    2019 doi

  12. [20]

    i wouldn’t say offensive but

    Vinitha Gadiraju, Shaun Kane, Sunipa Dev, Alex Taylor, Ding Wang, Emily Denton, and Robin Brewer. 2023. https://doi.org/10.1145/3593013.3593989 "i wouldn’t say offensive but...": Disability-centered perspectives on large language models . In Proceedings of the 2023 ACM Confere...

    2023

  13. [21]

    Leo Gao, Stella Biderman, Sid Black, Laurence Golding, Travis Hoppe, Charles Foster, Jason Phang, Horace He, Anish Thite, Noa Nabeshima, Shawn Presser, and Connor Leahy. 2020. https://arxiv.org/abs/2101.00027 The pile: An 800gb dataset of diverse text for language modeling . P...

    2020 arXiv

  14. [22]

    Vagrant Gautam, Arjun Subramonian, Anne Lauscher, and Os Keyes. 2024. https://doi.org/10.18653/v1/2024.gebnlp-1.20 Stop! in the name of flaws: Disentangling personal names and sociodemographic attributes in NLP . In Proceedings of the 5th Workshop on Gender Bias in Natural Lan...

    2024 doi

  15. [23]

    Kate Glazko, Yusuf Mohammed, Ben Kosa, Venkatesh Potluri, and Jennifer Mankoff. 2024. https://doi.org/10.1145/3630106.3658933 Identifying and improving disability bias in gpt-based resume screening . In Proceedings of the 2024 ACM Conference on Fairness, Accountability, and Tr...

    2024

  16. [24]

    Seraphina Goldfarb-Tarrant, Rebecca Marchant, Ricardo Mu \ n oz S \'a nchez, Mugdha Pandya, and Adam Lopez. 2021. https://doi.org/10.18653/v1/2021.acl-long.150 Intrinsic bias metrics do not correlate with application bias . In Proceedings of the 59th Annual Meeting of the Asso...

    2021 doi

  17. [25]

    Zishan Guo, Renren Jin, Chuang Liu, Yufei Huang, Dan Shi, Supryadi, Linhao Yu, Yan Liu, Jiaxuan Li, Bojian Xiong, and Deyi Xiong. 2023. https://arxiv.org/abs/2310.19736 Evaluating large language models: A comprehensive survey . Preprint, arXiv:2310.19736

    2023 arXiv

  18. [26]

    Amit Haim, Alejandro Salinas, and Julian Nyarko. 2024. https://arxiv.org/abs/2402.14875 What's in a name? auditing large language models for race and gender bias . Preprint, arXiv:2402.14875

    2024 arXiv

  19. [27]

    Hannah Herman. 2024. https://zapier.com/blog/automate-recruit-crm 6 ways to automate Recruit CRM with Zapier . Zapier Blog

    2024

  20. [28]

    Humanly. 2024. https://www.humanly.io/resource-library/blog/using-ai-to-streamline-the-recruiting-process Using AI to streamline the recruiting process . Humanly.io Blog

    2024

  21. [29]

    Albert Q. Jiang, Alexandre Sablayrolles, Antoine Roux, Arthur Mensch, Blanche Savary, Chris Bamford, Devendra Singh Chaplot, Diego de las Casas, Emma Bou Hanna, Florian Bressand, Gianna Lengyel, Guillaume Bour, Guillaume Lample, Lélio Renard Lavaud, Lucile Saulnier, Marie-Anne...

    2024 arXiv

  22. [30]

    Jack Kelly. 2023. https://www.forbes.com/sites/jackkelly/2023/03/15/how-ai-powered-tech-can-help-recruiters-and-hiring-managers-find-candidates-quicker-and-more-efficiently How AI -powered tech can help recruiters and hiring managers find candidates quicker and more efficientl...

    2023

  23. [31]

    JP Kincaid. 1975. Derivation of new readability formulas (automated readability index, fog count and flesch reading ease formula) for navy enlisted personnel. Chief of Naval Technical Training

    1975

  24. [32]

    Hannah Rose Kirk, Yennie Jun, Filippo Volpin, Haider Iqbal, Elias Benussi, Frederic Dreyer, Aleksandar Shtedritski, and Yuki Asano. 2021. https://proceedings.neurips.cc/paper_files/paper/2021/file/1531beb762df4029513ebf9295e0d34f-Paper.pdf Bias out-of-the-box: An empirical ana...

    2021

  25. [33]

    Matt Kusner, Joshua Loftus, Chris Russell, and Ricardo Silva. 2017. Counterfactual fairness. In Proceedings of the 31st International Conference on Neural Information Processing Systems, NIPS'17, page 4069–4079, Red Hook, NY, USA. Curran Associates Inc

    2017

  26. [34]

    u ttler, Mike Lewis, Wen-tau Yih, Tim Rockt\

    Patrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich K\" u ttler, Mike Lewis, Wen-tau Yih, Tim Rockt\" a schel, Sebastian Riedel, and Douwe Kiela. 2020. https://proceedings.neurips.cc/paper_files/paper/2020/file/6b493230205f78...

    2020

  27. [35]

    Haoyuan Li, Rui Zhang, and Snigdha Chaturvedi. 2025. https://aclanthology.org/2025.acl-short.90/ Improving fairness of large language models in multi-document summarization . In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 2: ...

    2025

  28. [36]

    Ninareh Mehrabi, Fred Morstatter, Nripsuta Saxena, Kristina Lerman, and Aram Galstyan. 2021. https://doi.org/10.1145/3457607 A survey on bias and fairness in machine learning . ACM Comput. Surv., 54(6)

    2021 doi

  29. [37]

    Meta AI . 2024. Meta llama 3. https://ai.meta.com/blog/meta-llama-3/. Blog post

    2024

  30. [38]

    Presidio Anonymizer

    Microsoft . Presidio Anonymizer . https://microsoft.github.io/presidio/anonymizer/

  31. [39]

    Huy Nghiem, John Prindle, Jieyu Zhao, and Hal Daum \'e Iii. 2024. https://aclanthology.org/2024.emnlp-main.413 `` you gotta be a doctor, lin '' : An investigation of name-based bias of large language models in employment recommendations . In Proceedings of the 2024 Conference ...

    2024

  32. [40]

    OpenAI . 2024. Gpt-4o. https://platform.openai.com/docs/models/gpt-4o

    2024

  33. [42]

    Marco Tulio Ribeiro, Tongshuang Wu, Carlos Guestrin, and Sameer Singh. 2020. https://doi.org/10.18653/v1/2020.acl-main.442 Beyond accuracy: Behavioral testing of NLP models with C heck L ist . In Proceedings of the 58th Annual Meeting of the Association for Computational Lingu...

    2020 doi

  34. [43]

    Javier S\' a nchez-Monedero, Lina Dencik, and Lilian Edwards. 2020. https://doi.org/10.1145/3351095.3372849 What does it mean to 'solve' the problem of discrimination in hiring? social, technical and legal perspectives from the uk on automated hiring systems . In Proceedings o...

    2020

  35. [44]

    Anurag Shandilya, Kripabandhu Ghosh, and Saptarshi Ghosh. 2018. https://doi.org/10.1145/3184558.3186947 Fairness of extractive text summarization . In Companion Proceedings of the The Web Conference 2018, WWW '18, page 97–98, Republic and Canton of Geneva, CHE. International W...

    2018

  36. [45]

    Emily Sheng, Kai-Wei Chang, Premkumar Natarajan, and Nanyun Peng. 2019. https://doi.org/10.18653/v1/D19-1339 The woman worked as a babysitter: On biases in language generation . In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9...

    2019 doi

  37. [46]

    Julius Steen and Katja Markert. 2024. https://arxiv.org/abs/2309.08047 Bias in news summarization: Measures, pitfalls and corpora . Preprint, arXiv:2309.08047

    2024 arXiv

  38. [47]

    Harini Suresh and John Guttag. 2021. https://doi.org/10.1145/3465416.3483305 A framework for understanding sources of harm throughout the machine learning life cycle . In Proceedings of the 1st ACM Conference on Equity and Access in Algorithms, Mechanisms, and Optimization, EA...

    2021

  39. [48]

    Alex Tamkin, Amanda Askell, Liane Lovitt, Esin Durmus, Nicholas Joseph, Shauna Kravec, Karina Nguyen, Jared Kaplan, and Deep Ganguli. 2023. https://arxiv.org/abs/2312.03689 Evaluating and mitigating discrimination in language model decisions . Preprint, arXiv:2312.03689

    2023 arXiv

  40. [49]

    Yixin Wan, George Pu, Jiao Sun, Aparna Garimella, Kai-Wei Chang, and Nanyun Peng. 2023. https://doi.org/10.18653/v1/2023.findings-emnlp.243 `` kelly is a warm person, joseph is a role model '' : Gender biases in LLM -generated reference letters . In Findings of the Association...

    2023 doi

  41. [50]

    Yuan Wang, Xuyang Wu, Hsin-Tai Wu, Zhiqiang Tao, and Yi Fang. 2024. https://doi.org/10.18653/v1/2024.naacl-long.319 Do large language models rank fairly? an empirical study on the fairness of LLM s as rankers . In Proceedings of the 2024 Conference of the North American Chapte...

    2024 doi

  42. [51]

    Kellie Webster, Xuezhi Wang, Ian Tenney, Alex Beutel, Emily Pitler, Ellie Pavlick, Jilin Chen, Ed Chi, and Slav Petrov. 2021. https://arxiv.org/abs/2010.06032 Measuring and reducing gendered correlations in pre-trained models . Preprint, arXiv:2010.06032

    2021 arXiv

  43. [52]

    Kyra Wilson and Aylin Caliskan. 2024. https://arxiv.org/abs/2407.20371 Gender, race, and intersectional bias in resume screening via language model retrieval . Preprint, arXiv:2407.20371

    2024 arXiv

  44. [53]

    Chen Xu, Wenjie Wang, Yuxin Li, Liang Pang, Jun Xu, and Tat-Seng Chua. 2024. https://doi.org/10.18653/v1/2024.findings-emnlp.467 A study of implicit ranking unfairness in large language models . In Findings of the Association for Computational Linguistics: EMNLP 2024, pages 79...

    2024 doi

  45. [54]

    Leon Yin, Davey Alba, and Leonardo Nicoletti. 2024. https://www.bloomberg.com/graphics/2024-openai-gpt-hiring-racial-discrimination/ Openai’s gpt is a recruiter’s dream tool. tests show there’s racial bias . Bloomberg

    2024

  46. [55]

    Yusen Zhang, Nan Zhang, Yixin Liu, Alexander Fabbri, Junru Liu, Ryo Kamoi, Xiaoxin Lu, Caiming Xiong, Jieyu Zhao, Dragomir Radev, Kathleen McKeown, and Rui Zhang. 2024. https://doi.org/10.18653/v1/2024.naacl-long.187 Fair abstractive summarization of diverse perspectives . In ...

    2024 doi

  47. [56]

    Jieyu Zhao, Tianlu Wang, Mark Yatskar, Vicente Ordonez, and Kai-Wei Chang. 2018. https://doi.org/10.18653/v1/N18-2003 Gender bias in coreference resolution: Evaluation and debiasing methods . In Proceedings of the 2018 Conference of the North A merican Chapter of the Associati...

    2018 doi

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.