Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

Fair-PP: A Synthetic Dataset for Aligning LLM with Personalized Preferences of Social Equity

T0 review · 4 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read Synthetic role-play data lets a small LLM align to a target persona's equity preferences while staying distinct from other personas.

desk verdict A useful synthetic dataset and a sensible reweighting method, but the alignment claims only reach simulated personas until human validation is added. read the letter →

arxiv 2505.11861 v1 pith:BC2HAZN2 submitted 2025-05-17 cs.AI cs.CL

Qi Zhou , Jie Zhang , Dongxia Wang , Qiang Liu , Tianlin Li , Jin Song Dong , Wenhai Wang , Qing Guo This is my paper · ORCID
classification cs.AIcs.CL
keywords personalizedpreferencessocialequitysyntheticdatasetLLMalignmentdirectpreferenceoptimizationpersonarole-playsamplereweightingJensen-Shannondistance
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

FAIR-PP is a synthetic dataset of 238,623 personalized social-equity preference records, generated by having GPT-4o-mini role-play seven personas drawn from a real UK social survey and answer questions built from 28 social groups, 98 equity topics, and five value dimensions. The paper's central claim is that this dataset lets a small language model be fine-tuned to match a target persona's equity preferences while becoming measurably more distinct from the other six personas. Its sample reweighting method, applied to direct preference optimization, is reported to reach 0.98 similarity to the target persona on held-out questions while increasing separation from other personas by 10.20 to 12.00 percent over vanilla DPO. If true, the result matters because personalized value alignment could be produced cheaply and automatically, without collecting new human preference labels.

What carries the argument

The central mechanism is a persona-anchored preference space. The seven persona portrayals are fixed by their answer distributions over 34,089 questions, and any model's answer distribution is located in the space by 1 minus Jensen-Shannon distance to each persona anchor. The alignment mechanism is the sample reweighting formula $W_i = (T_i/N_i)/\sum_j T_j/N_j$, which upweights training examples where the target persona is in a small agreement tier and downweights common answers; the reweighted samples are then fed to SFT or DPO. This weighting is what converts a generic alignment step into a personalization step that separates the target persona from the others.

What would settle it

Give the FAIR-PP questionnaire to a representative sample of real people in each of the seven UK segments and compare their answer distributions with GPT-4o-mini's role-play responses; if the Jensen-Shannon distance between human and simulated distributions is not near zero on a held-out question set, the dataset's labels do not stand in for human preferences.

Watch

Extended reading notes

Core claim

The paper builds FAIR-PP by combining 28 social groups, 98 equity topics, and five binary perspective dimensions (meritocracy versus egalitarianism, procedural versus distributive justice, individualism versus collectivism, social norm versus equity concerns, moral versus legal), producing multiple-choice questions with two value-opposed options and an N/A option. A commercial LLM, GPT-4o-mini, answers every question under seven persona prompts derived from a 2020 UK survey, yielding 238,623 preference records. The paper treats the seven personas as anchor points in a personalized preference space, and positions six open-weight regional LLMs inside that space; all six sit closest to Persona 3, the Disengaged Battlers. For alignment, it proposes sample-level reweighting: answers on which the target persona agrees with many other personas are down-weighted, and answers unique to the target persona are up-weighted, with weights set by the formula $W_i = (T_i/N_i)/\sum_j T_j/N_j$ where $T_i$ is the agreement-frequency tier and $N_i$ the count in that tier. Fine-tuning Llama-3.2-3B with weighted DPO raises similarity to target Persona 6 to 0.98 on held-out questions while increasing average distance from the other personas by 10.20 to 12.00 percent over vanilla DPO; on generated scenario variants it reaches 0.87 similarity and cuts similarity to other personas by 37.80 percent relative to vanilla.

Load-bearing premise

The load-bearing premise is that a language model instructed to role-play a persona produces equity preferences matching those of real people in that social segment, so the synthetic labels are treated as ground truth for human preferences.

Editorial extensions

If this is right

  • Personalized social-equity alignment can be done from synthetic role-play data alone, avoiding the cost of collecting new human preference labels.
  • Weighting training samples by how distinctive they are for the target persona improves both alignment to the target and separation from the other personas; weighted DPO is the strongest of the tested methods.
  • Mainstream open-weight LLMs are not value-neutral in the FAIR-PP space: all six tested models are nearest to the Disengaged Battlers persona, Persona 3.
  • Persona alignment transfers to generated scenario variants of the test questions, indicating the fine-tuning is not merely memorizing template wording.
  • The same resource can serve both as a benchmark for mapping LLM value positions and as training data for re-targeting those positions.

Reading between the lines

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

  • Since evaluation uses the same synthetic persona space that generated the labels, the reported gains show consistency with simulated personas; a human validation study is needed before claiming alignment with real human values.
  • The uniqueness weighting treats rarity as persona identity, so it may overfit to idiosyncratic or low-quality questions; holding out entire equity topics would test whether the persona profile is coherent rather than memorized.
  • The same role-play-plus-reweighting recipe could extend to other contested value domains, but the paper only demonstrates it for social equity with personas from one country and one time point.
  • Mapping non-UK models against UK-derived persona anchors may reflect the anchor set more than actual regional value differences.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The paper introduces FAIR-PP, a synthetic dataset of 238,623 personalized preference records on social equity, generated by having GPT-4o-mini role-play seven personas derived from the 2020 UK More in Common survey. The dataset is built from a question bank covering 28 social groups, 98 equity topics, and five preference dimensions. The authors also contribute an automated data generation framework, an analysis of where six mainstream LLMs sit in the resulting persona-anchored preference space, and a sample-reweighting method (WSFT/WDPO) for aligning a model with a target persona while increasing divergence from other personas. Experiments on Llama-3.2-3B report that WDPO achieves 0.98 similarity to Persona 6 on the held-out original test set and 0.87 on the generation-based simulation set, outperforming DPO and SFT baselines.

Significance. If the results hold, FAIR-PP is a useful resource for studying persona-conditioned preference alignment at scale, and the reweighting idea—upweighting preference samples that are unique to a target persona—is a simple, plausible mechanism for increasing alignment specificity. The public release of the dataset and code is a concrete strength, and the automated generation framework could be adapted to other survey-derived value dimensions. The main significance is qualified by a fundamental external-validity gap: all training labels and evaluation anchors come from the same GPT-4o-mini role-play process, so the reported alignment gains have not been shown to transfer to actual human segments. The paper's contribution is better described as 'alignment with simulated personas' unless independent human validation is added.

major comments (4)
  1. [Section 2.5 and Sections 3.2–3.3] The training preference labels and the evaluation anchors are both produced by GPT-4o-mini role-playing the same seven persona descriptions defined in Section 2.5. The held-out test set is a random split of the same synthetic data, and the simulation set in Section 3.3 is generated by GPT-4o from the test questions. Consequently, the 0.98 similarity to Persona 6 reported in Table 2 measures agreement with the label generator's role-play distribution, not with any real human segment. The paper should either add an independent human validation subset (e.g., real survey responses matched to the seven UK segments) or systematically reframe all 'personalized preferences' claims as 'simulated persona preferences' throughout the abstract, introduction, and conclusions.
  2. [Tables 2 and 3] No error bars, multiple seeds, or significance tests are reported. The central comparison between WDPO and DPO is numerically identical on the original test set for Persona 6 (0.98 vs. 0.98 in Table 2), and the simulation-set difference (0.87 vs. 0.77 in Table 3) could fall within run-to-run noise. The authors should report mean and standard deviation over at least three to five random seeds and include a paired significance test (e.g., bootstrap or Wilcoxon) for the differences that support the central claim.
  3. [Section 3.2 and Equation (1)] The reported 'margin' improvements of 10.20% and 12.00% are undefined, and the phrase 'compared to vanilla' is ambiguous (vanilla DPO or unaligned vanilla?). Please define a scalar divergence metric, for example the average 1 - Jensen-Shannon distance to all non-target personas, and state the baseline explicitly. In addition, the reweighting formula depends on the arbitrary mapping from matching-frequency counts to tier numbers in Appendix E (Table 4); report the sensitivity of the results to reasonable alternative mappings.
  4. [Section 3.1 and Table 1] The claim of positioning LLMs 'across five major global regions' is not supported by the experimental design. Each region is represented by at most one model (two models for North America), so the observed similarities, such as Qwen2.5-7B being closest to Persona 3, conflate model architecture/company with regional representation. To support a regional claim, the authors should either include multiple models per region or explicitly temper the contribution to 'positioning of several representative open-weight models.'
minor comments (5)
  1. [Section 2.2 and Abstract] The text says the topics total 97 specific topics, while the abstract and Figure 2 report 98 equity topics. Please reconcile the count.
  2. [Section 2.6] The cross-reference 'In the subsequent Section 2.1' appears to intend Section 3.1, since Section 2.1 is about social groups. Please correct the reference.
  3. [Section 3.3] The statement that WDPO 'decrease[s] 37.80% compared to vanilla' does not specify the baseline (vanilla DPO or unaligned vanilla) or the metric used. Please define both.
  4. [Appendix E] There is a formatting issue: 'batch size to32and16' should read 'batch size to 32 and 16'. Please also define all hyperparameters for the reweighting tiers in a single table.
  5. [Abstract] The abstract mentions 'five major global regions,' but Section 3.1 evaluates six models from five regions. Please clarify whether the analysis is at the model level or the region level.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the synthetic dataset and alignment evaluation form an internally consistent benchmark, with human-validation absence as a correctness limitation, not a derivation loop.

full rationale

I find no step in the claimed derivation chain that reduces to its own inputs by construction. FAIR-PP is explicitly introduced as a synthetic dataset: Section 2.5 states that 'we leverage GPT-4o-mini to simulate seven representative personas to capture the diverse personalized preference,' and the persona anchors used for evaluation are constructed from the same role-play procedure. The alignment experiments in Sections 3.2 and 3.3 split this synthetic dataset into held-out training and test portions and measure 1 - Jensen-Shannon distance to those anchors. That is a self-contained internal benchmark rather than a circular derivation: the reported scores (e.g., 0.98 toward Persona 6) are empirical generalization results on held-out synthetic labels, not quantities that equal the training objective by definition. The sample reweighting formula in Equation (1) is a method applied to the dataset, and its reported gains are relative comparisons against vanilla SFT and DPO on the same held-out data. There are no load-bearing self-citations; the cited works for the similarity metric and self-calibration prompt are external and not authored by the present paper's authors. The paper's own Section 5 limitation—'incorporating human survey responses or validation could further enhance data quality'—correctly identifies the absence of human validation. This is an external-validity concern about whether the synthetic labels reflect real human equity preferences, not a circularity in the paper's internal derivation. The central claims are honestly scoped to a synthetic preference space, so the circularity score is 0.

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

The central claim rests on an LLM-generated synthetic ground truth. The free parameters are the hand-chosen tier mapping in the reweighting scheme; the axioms include role-play fidelity, cross-regional transfer of UK personas, and the adequacy of the five equity dimensions. No new physical or conceptual entities are introduced beyond the persona anchors, which are derived from a real survey.

free parameters (1)
  • Frequency tier mapping for sample reweighting = Tier mapping 6->1, 5->2, 4->3, 3->4, 2->5, 1->6, 0->7; weights 0.015 to 0.205
    Appendix E and Equation 1. The ordinal mapping from matching counts to tiers and the resulting weights are hand-chosen by the authors and depend on the training split; the claimed 10.20% and 12.00% divergence improvements are sensitive to this mapping.
assumptions (4)
  • domain assumption LLM role-play with self-calibration produces valid proxies for real human personas
    Section 2.5 uses GPT-4o-mini to generate all preference labels; no human survey data validates the role-play outputs.
  • domain assumption The seven UK More in Common personas generalize to global regions
    Section 3.1 positions models from five regions against these UK-derived personas; the paper anonymizes country-specific details but does not establish cross-cultural validity.
  • domain assumption The five preference dimensions span the relevant equity preference space
    Section 2.3 defines five dimensions (meritocracy/egalitarianism, procedural/distributive, individualism/collectivism, social norm/equity, moral/law); the choice is grounded in cited literature but is an author-selected taxonomy.
  • standard math Jensen-Shannon distance is an appropriate similarity metric for preference distributions
    Used in Section 2.6 and experiments; follows prior work [28].

how reviews work

0 comments
Cite this review

Pith. "Pith review of Fair-PP: A Synthetic Dataset for Aligning LLM with Personalized Preferences of Social Equity." pith.science (2026). https://pith.science/paper/BC2HAZN2

@misc{pith2026250511861,
  author       = {Pith},
  title        = {Pith review of: Fair-PP: A Synthetic Dataset for Aligning LLM with Personalized Preferences of Social Equity},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/BC2HAZN2}},
  note         = {Machine review of arXiv:2505.11861}
}
read the original abstract

Human preference plays a crucial role in the refinement of large language models (LLMs). However, collecting human preference feedback is costly and most existing datasets neglect the correlation between personalization and preferences. To address this issue, we introduce Fair-PP, a synthetic dataset of personalized preferences targeting social equity, derived from real-world social survey data, which includes 28 social groups, 98 equity topics, and 5 personal preference dimensions. Leveraging GPT-4o-mini, we engage in role-playing based on seven representative persona portrayals guided by existing social survey data, yielding a total of 238,623 preference records. Through Fair-PP, we also contribute (i) An automated framework for generating preference data, along with a more fine-grained dataset of personalized preferences; (ii) analysis of the positioning of the existing mainstream LLMs across five major global regions within the personalized preference space; and (iii) a sample reweighting method for personalized preference alignment, enabling alignment with a target persona while maximizing the divergence from other personas. Empirical experiments show our method outperforms the baselines.

Figures

Figures reproduced from arXiv: 2505.11861 by the authors.

Figure 1
Figure 1. Landscape of FAIR-PP space. Despite increasing efforts in exploring and collecting hu￾man preference data of social values, existing datasets face several limitations: (i) Social equity unaddressed: lack consideration of social equity which is an important is￾sue in social psychology5 (ii) Neglecting personalization: Most of them emphasize universal viewpoints, e.g., cul￾tural and political beliefs [10] lacking cons… view at source ↗
Figure 2
Figure 2. An overview of the FAIR-PP dataset. Each question consists of three parts: the social group, an equity topic, and a perspective dimension. An example question is shown on the right where option A and B represent two types of viewpoints under a specific dimension. Personalized preferences are collected through LLM personalization that leverages 7 value portrait based on the real-world social surveys. 2 FAIR-PP FAIR-P… view at source ↗
Figure 3
Figure 3. Personalized preference anchors. Blue and red represent the proportions of choices for [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Fine-grained Personalized Preferences and Aggregate Distribution Across Personas, Social [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Personalized preference analysis: (a) we calculate the Jensen-Shannon similarity between [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: Question data point example. • Persona 3 (Disengaged Battlers): Pleases act as one of Disengaged Battlers, you are focused on the everyday struggle for survival. You have work, but often it is insecure or involves irregular hours. You tend to feel disconnected from oth…
Figure 7
Figure 7. Figure 7: Heatmap across social groups [PITH_FULL_IMAGE:figures/full_fig_p018_7.png]
Figure 8
Figure 8. Figure 8: Heatmap across equity topics. E Experimental Details As for frequency tier, we mapping different matching counts from 0 to 6 to tier numbers from 7 to 1, respectively. For instance, the weights corresponding to different frequency tiers of persona 6 are as [PITH_FULL_…

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. Reward Shaping for (Inference-Time) Alignment: A Stackelberg Game Perspective

    cs.LG 2026-01 conditional novelty 5.0 of 10

    The optimal reward for KL-regularized LLM alignment is a threshold function—reward B above a prompt-dependent cutoff, 0 below—which can be estimated from base-model samples and integrated into decoding-time alignment.

Reference graph

Works this paper leans on

42 extracted references · 18 canonical work pages · cited by 1 Pith paper

  1. [1]

    A new perspective on equity theory: The equity sensitivity construct.Academy of management Review, 12(2):222–234, 1987

    Richard C Huseman, John D Hatfield, and Edward W Miles. A new perspective on equity theory: The equity sensitivity construct.Academy of management Review, 12(2):222–234, 1987

    work page 1987

  2. [2]

    Understanding equity sensitivity through the lens of personality: a review of associations and underlying nature.Management Research Review, 46(9):1261–1277, 2023

    Nikhita Tuli, Kunal Shrivastava, and Disha Khattar. Understanding equity sensitivity through the lens of personality: a review of associations and underlying nature.Management Research Review, 46(9):1261–1277, 2023

    work page 2023

  3. [3]

    Beavertails: Towards improved safety alignment of llm via a human-preference dataset.Advances in Neural Information Processing Systems, 36:24678–24704, 2023

    Jiaming Ji, Mickel Liu, Josef Dai, Xuehai Pan, Chi Zhang, Ce Bian, Boyuan Chen, Ruiyang Sun, Yizhou Wang, and Yaodong Yang. Beavertails: Towards improved safety alignment of llm via a human-preference dataset.Advances in Neural Information Processing Systems, 36:24678–24704, 2023

    2023

  4. [4]

    Safety alignment should be made more than just a few tokens deep.arXiv preprint arXiv:2406.05946, 2024

    Xiangyu Qi, Ashwinee Panda, Kaifeng Lyu, Xiao Ma, Subhrajit Roy, Ahmad Beirami, Prateek Mittal, and Peter Henderson. Safety alignment should be made more than just a few tokens deep.arXiv preprint arXiv:2406.05946, 2024

    arXiv 2024

  5. [5]

    Safeworld: Geo-diverse safety alignment.Advances in Neural Information Processing Systems, 37:128734– 128768, 2024

    Da Yin, Haoyi Qiu, Kung-Hsiang Huang, Kai-Wei Chang, and Nanyun Peng. Safeworld: Geo-diverse safety alignment.Advances in Neural Information Processing Systems, 37:128734– 128768, 2024

    work page 2024

  6. [6]

    Lisa: Lazy safety alignment for large language models against harmful fine-tuning attack.Advances in Neural Information Processing Systems, 37:104521–104555, 2024

    Tiansheng Huang, Sihao Hu, Fatih Ilhan, Selim Tekin, and Ling Liu. Lisa: Lazy safety alignment for large language models against harmful fine-tuning attack.Advances in Neural Information Processing Systems, 37:104521–104555, 2024

    2024

  7. [7]

    Cultural bias and cultural alignment of large language models.PNAS nexus, 3(9):pgae346, 2024

    Yan Tao, Olga Viberg, Ryan S Baker, and René F Kizilcec. Cultural bias and cultural alignment of large language models.PNAS nexus, 3(9):pgae346, 2024

    2024

  8. [8]

    Investigating cultural alignment of large language models

    Badr Alkhamissi, Muhammad ElNokrashy, Mai Alkhamissi, and Mona Diab. Investigating cultural alignment of large language models. InProceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 12404–12422, 2024

    work page 2024

Show all 42 references
  1. [9]

    Culturellm: Incorporating cultural differences into large language models.Advances in Neural Information Processing Systems, 37:84799–84838, 2024

    Cheng Li, Mengzhuo Chen, Jindong Wang, Sunayana Sitaram, and Xing Xie. Culturellm: Incorporating cultural differences into large language models.Advances in Neural Information Processing Systems, 37:84799–84838, 2024

    2024

  2. [10]

    Cul- turepark: Boosting cross-cultural understanding in large language models.arXiv preprint arXiv:2405.15145, 2024

    Cheng Li, Damien Teney, Linyi Yang, Qingsong Wen, Xing Xie, and Jindong Wang. Cul- turepark: Boosting cross-cultural understanding in large language models.arXiv preprint arXiv:2405.15145, 2024

    2024 arXiv

  3. [11]

    Towards measuring the representation of subjective global opinions in language models.arXiv preprint arXiv:2306.16388, 2023

    Esin Durmus, Karina Nguyen, Thomas I Liao, Nicholas Schiefer, Amanda Askell, Anton Bakhtin, Carol Chen, Zac Hatfield-Dodds, Danny Hernandez, Nicholas Joseph, et al. Towards measuring the representation of subjective global opinions in language models.arXiv preprint arXiv:2306....

    2023 arXiv

  4. [12]

    Whose opinions do language models reflect? InInternational Conference on Machine Learning, pages 29971–30004

    Shibani Santurkar, Esin Durmus, Faisal Ladhak, Cinoo Lee, Percy Liang, and Tatsunori Hashimoto. Whose opinions do language models reflect? InInternational Conference on Machine Learning, pages 29971–30004. PMLR, 2023

    2023

  5. [13]

    Value kaleidoscope: Engaging ai with pluralistic human values, rights, and duties

    Taylor Sorensen, Liwei Jiang, Jena D Hwang, Sydney Levine, Valentina Pyatkin, Peter West, Nouha Dziri, Ximing Lu, Kavel Rao, Chandra Bhagavatula, et al. Value kaleidoscope: Engaging ai with pluralistic human values, rights, and duties. InProceedings of the AAAI Conference on A...

    2024

  6. [14]

    Worldvaluesbench: A large-scale benchmark dataset for multi-cultural value awareness of language models

    Wenlong Zhao, Debanjan Mondal, Niket Tandon, Danica Dillion, Kurt Gray, and Yuling Gu. Worldvaluesbench: A large-scale benchmark dataset for multi-cultural value awareness of language models. InProceedings of the 2024 Joint International Conference on Computational Linguistics...

    2024

  7. [15]

    Alpacafarm: A simulation framework for methods that learn from human feedback.Advances in Neural Information Processing Systems, 36:30039–30069, 2023

    Yann Dubois, Chen Xuechen Li, Rohan Taori, Tianyi Zhang, Ishaan Gulrajani, Jimmy Ba, Carlos Guestrin, Percy S Liang, and Tatsunori B Hashimoto. Alpacafarm: A simulation framework for methods that learn from human feedback.Advances in Neural Information Processing Systems, 36:3...

    2023

  8. [16]

    Ultrafeedback: Boosting language models with scaled ai feedback.arXiv preprint arXiv:2310.01377, 2023

    Ganqu Cui, Lifan Yuan, Ning Ding, Guanming Yao, Bingxiang He, Wei Zhu, Yuan Ni, Guotong Xie, Ruobing Xie, Yankai Lin, et al. Ultrafeedback: Boosting language models with scaled ai feedback.arXiv preprint arXiv:2310.01377, 2023

    arXiv

  9. [17]

    Social change, cultural evolution, and human development.Current Opinion in Psychology, 8:84–92, 2016

    Patricia M Greenfield. Social change, cultural evolution, and human development.Current Opinion in Psychology, 8:84–92, 2016

    2016

  10. [18]

    Humans adapt to social diversity over time.Proceedings of the National Academy of Sciences, 116(25):12244– 12249, 2019

    Miguel R Ramos, Matthew R Bennett, Douglas S Massey, and Miles Hewstone. Humans adapt to social diversity over time.Proceedings of the National Academy of Sciences, 116(25):12244– 12249, 2019

    2019

  11. [19]

    Modeling and forecasting the evolution of prefer- ences over time: A hidden markov model of travel behavior.arXiv preprint arXiv:1707.09133, 2017

    Feras El Zarwi, Akshay Vij, and Joan Walker. Modeling and forecasting the evolution of prefer- ences over time: A hidden markov model of travel behavior.arXiv preprint arXiv:1707.09133, 2017

    2017 arXiv

  12. [20]

    The measurement of equity sensitivity.Journal of Occupational and Organizational Psychology, 67(2):133–142, 1994

    Wesley C King Jr and Edward W Miles. The measurement of equity sensitivity.Journal of Occupational and Organizational Psychology, 67(2):133–142, 1994

    1994

  13. [21]

    Britain’s choice: Polarisation or cohesion.The Political Quarterly, 92(1):119– 124, 2021

    Paula Surridge. Britain’s choice: Polarisation or cohesion.The Political Quarterly, 92(1):119– 124, 2021

    2021

  14. [22]

    Belief in egalitarianism and meritocracy.Economics Letters, 221:110896, 2022

    Hideaki Goto. Belief in egalitarianism and meritocracy.Economics Letters, 221:110896, 2022

    2022

  15. [23]

    Procedural justice, distributive justice: How experiences with downsizing condition their impact on organizational commitment.Social Psychology Quarterly, 68(1):89–102, 2005

    Jody Clay-Warner, Karen A Hegtvedt, and Paul Roman. Procedural justice, distributive justice: How experiences with downsizing condition their impact on organizational commitment.Social Psychology Quarterly, 68(1):89–102, 2005

    2005

  16. [24]

    Culture matters: Individualism vs

    Rebecca LeFebvre and V olker Franke. Culture matters: Individualism vs. collectivism in conflict decision-making.Societies, 3(1):128–146, 2013

    2013

  17. [25]

    Social norms and gender disparities with a focus on female labor force participation in south asia.The World Bank Research Observer, 39(1):124– 158, 2024

    Maurizio Bussolo, Jessy Amarachi Ezebuihe, Ana Maria Muñoz Boudet, Stavros Poupakis, Tasmia Rahman, and Nayantara Sarma. Social norms and gender disparities with a focus on female labor force participation in south asia.The World Bank Research Observer, 39(1):124– 158, 2024

    2024

  18. [26]

    Achieving ethics and fairness in hiring: Going beyond the law.Journal of Business Ethics, 68:449–464, 2006

    G Stoney Alder and Joseph Gilbert. Achieving ethics and fairness in hiring: Going beyond the law.Journal of Business Ethics, 68:449–464, 2006

    2006

  19. [27]

    Gpt-4 technical report.arXiv preprint arXiv:2303.08774, 2023

    Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Florencia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, et al. Gpt-4 technical report.arXiv preprint arXiv:2303.08774, 2023

    2023 arXiv

  20. [28]

    Modular pluralism: Pluralistic alignment via multi-llm collaboration.arXiv preprint arXiv:2406.15951, 2024

    Shangbin Feng, Taylor Sorensen, Yuhan Liu, Jillian Fisher, Chan Young Park, Yejin Choi, and Yulia Tsvetkov. Modular pluralism: Pluralistic alignment via multi-llm collaboration.arXiv preprint arXiv:2406.15951, 2024

    2024 arXiv

  21. [29]

    Falcon 3 family of open foundation models, December 2024

    TII Team. Falcon 3 family of open foundation models, December 2024

    2024

  22. [30]

    The llama 3 herd of models.arXiv preprint arXiv:2407.21783, 2024

    Aaron Grattafiori, Abhimanyu Dubey, Abhinav Jauhri, Abhinav Pandey, Abhishek Kadian, Ahmad Al-Dahle, Aiesha Letman, Akhil Mathur, Alan Schelten, Alex Vaughan, et al. The llama 3 herd of models.arXiv preprint arXiv:2407.21783, 2024

    2024 arXiv

  23. [31]

    Sea-lion: Southeast asian languages in one network.arXiv preprint arXiv:2504.05747, 2025

    Raymond Ng, Thanh Ngan Nguyen, Yuli Huang, Ngee Chia Tai, Wai Yi Leong, Wei Qi Leong, Xianbin Yong, Jian Gang Ngui, Yosephine Susanto, Nicholas Cheng, et al. Sea-lion: Southeast asian languages in one network.arXiv preprint arXiv:2504.05747, 2025

    2025

  24. [32]

    Albert Q Jiang, Alexandre Sablayrolles, Arthur Mensch, Chris Bamford, and Devendra Singh Chaplot. Diego de las casas.Florian Bressand, Gianna Lengyel, Guillaume Lample, Lucile Saulnier, Lélio Renard Lavaud, Marie-Anne Lachaux, Pierre Stock, Teven Le Scao, Thibaut Lavril, Thoma...

    2023

  25. [33]

    An Yang, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chengyuan Li, Dayiheng Liu, Fei Huang, Haoran Wei, et al. Qwen2. 5 technical report.arXiv preprint arXiv:2412.15115, 2024

    2024 arXiv

  26. [34]

    Direct preference optimization: Your language model is secretly a reward model

    Rafael Rafailov, Archit Sharma, Eric Mitchell, Christopher D Manning, Stefano Ermon, and Chelsea Finn. Direct preference optimization: Your language model is secretly a reward model. Advances in Neural Information Processing Systems, 36:53728–53741, 2023

    2023

  27. [35]

    Deep reinforcement learning from human preferences.Advances in neural information processing systems, 30, 2017

    Paul F Christiano, Jan Leike, Tom Brown, Miljan Martic, Shane Legg, and Dario Amodei. Deep reinforcement learning from human preferences.Advances in neural information processing systems, 30, 2017

    2017

  28. [36]

    Fine-tuning language models from human preferences

    Daniel M Ziegler, Nisan Stiennon, Jeffrey Wu, Tom B Brown, Alec Radford, Dario Amodei, Paul Christiano, and Geoffrey Irving. Fine-tuning language models from human preferences. arXiv preprint arXiv:1909.08593, 2019

    1909 arXiv

  29. [37]

    Training a helpful and harmless assistant with reinforcement learning from human feedback.arXiv preprint arXiv:2204.05862, 2022

    Yuntao Bai, Andy Jones, Kamal Ndousse, Amanda Askell, Anna Chen, Nova DasSarma, Dawn Drain, Stanislav Fort, Deep Ganguli, Tom Henighan, et al. Training a helpful and harmless assistant with reinforcement learning from human feedback.arXiv preprint arXiv:2204.05862, 2022

    2022 arXiv

  30. [38]

    Normad: A framework for measuring the cultural adaptability of large language models.arXiv preprint arXiv:2404.12464, 2024

    Abhinav Rao, Akhila Yerukola, Vishwa Shah, Katharina Reinecke, and Maarten Sap. Normad: A framework for measuring the cultural adaptability of large language models.arXiv preprint arXiv:2404.12464, 2024

    2024 arXiv

  31. [39]

    Hannah Rose Kirk, Alexander Whitefield, Paul Rottger, Andrew M Bean, Katerina Margatina, Rafael Mosquera-Gomez, Juan Ciro, Max Bartolo, Adina Williams, He He, et al. The prism alignment dataset: What participatory, representative and individualised human feedback reveals about...

    2024

  32. [40]

    Group robust preference optimization in reward-free rlhf.Advances in Neural Information Processing Systems, 37:37100–37137, 2024

    Shyam Sundhar Ramesh, Yifan Hu, Iason Chaimalas, Viraj Mehta, Pier Giuseppe Sessa, Haitham Bou Ammar, and Ilija Bogunovic. Group robust preference optimization in reward-free rlhf.Advances in Neural Information Processing Systems, 37:37100–37137, 2024

    2024

  33. [41]

    Interpretable preferences via multi-objective reward modeling and mixture-of-experts.arXiv preprint arXiv:2406.12845, 2024

    Haoxiang Wang, Wei Xiong, Tengyang Xie, Han Zhao, and Tong Zhang. Interpretable preferences via multi-objective reward modeling and mixture-of-experts.arXiv preprint arXiv:2406.12845, 2024

    2024 arXiv

  34. [42]

    Whose boat does it float? improving personalization in preference tuning via inferred user personas.arXiv preprint arXiv:2501.11549, 2025

    Nishant Balepur, Vishakh Padmakumar, Fumeng Yang, Shi Feng, Rachel Rudinger, and Jor- dan Lee Boyd-Graber. Whose boat does it float? improving personalization in preference tuning via inferred user personas.arXiv preprint arXiv:2501.11549, 2025. A Details about social groups a...

    2025 arXiv

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.