The study examines how large language models (LLMs) decide who gets access to restricted academic resources in a controlled simulation. Models consistently favor PhD candidates, but their preferences shift with the requester's region and academic status. Large proprietary LLMs tend to favor low‑income or Global South applicants, while smaller open‑weight models favor the Global North, suggesting that model scale, alignment, and training data influence gatekeeping biases.
arXiv:2608. 05178v1 Announce Type: cross Abstract: Equitable access to scientific knowledge often depends on informal gatekeeping decisions, particularly when resources such as paywalled articles, datasets, or professional materials such as curriculum vitae (CV) must be shared selectively.
arXiv:2607. 28934v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly involved in the distribution of scarce resources, raising concerns about biased allocations based on characteristics like race and gender.
The study audits demographic bias across four deep knowledge tracing architectures—DKT, DKVMN, SAKT, and AKT—using two large public datasets (Eedi and OULAD). It finds that bias is context‑dependent: socioeconomic bias is significant on Eedi, while gender bias appears on OULAD for most models. The most accurate model, AKT, also exhibits the greatest bias, and standard mitigation techniques such as reweighting and adversarial debiasing fail to reduce bias without sacrificing accuracy.
The paper investigates how large language models (LLMs) from OpenAI and Anthropic exhibit probability bias when deciding whether to trust information from reporters of different racial descriptions. Using a two‑urn Bayesian game and 6,600 sessions, the authors find that models often refuse to assess a reporter’s credibility—especially for Black‑described reporters—yet when they do assess, Black reporters receive slightly higher trust. The bias is subtle, growing with reasoning effort, and is not detected by standard audits that compare average trust or only assessed cases.
The paper introduces Thomson, a frontier AI model developed through continual learning on open-weight models, aiming to democratize access to high-performance AI. It argues that institutions with limited resources can achieve frontier-level performance by applying a modern mid- & post-training stack, preserving model plasticity and stability while minimizing high-impact interventions. Thomson demonstrates competitive performance across agentic tasks, safety, legal, tax, multilingualism, and large-scale deep research, exhibiting a distinctive π-shaped improvement pattern and effectively mitigating the forgetting problem seen in narrow domain adaptation.
The study evaluates ten bias audit instruments across ten advanced language models on occupational gender, age, and socioeconomic status. While each tool reliably detects bias, their rankings of model performance are essentially random, indicating that different audits measure distinct constructs. The findings show that a single audit can identify bias direction within its own framework, but no audit can consistently rank models against one another.
arXiv:2606. 26099v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly deployed in artificial intelligence (AI) governance analysis across national and international organisations.
arXiv:2608. 07762v1 Announce Type: new Abstract: LLM benchmarks can build an organization's reputation and attract customers, but only when results are transparent and verifiable.
The study examines how the choice of large language model (LLM) used to write job application cover letters influences hiring decisions, independent of the applicant’s qualifications. Using a multi‑model audit with eleven writer agents and nine evaluator agents across ten job descriptions, the authors find a consistent ranking of writer models—locally hosted models are disadvantaged and evaluators often favor their own outputs. The hierarchy has tangible effects, with up to 1.8 of 25 qualified candidates being rejected solely due to model mismatch, and the patterns are explained by shared stylistic cues that both writers and evaluators reward or penalise.
arXiv:2608.02551v2 Announce Type: replace-cross Abstract: Fairness evaluation concerns not only what a model produces, but also what its outputs ought to be compared against. When a model generates "...
arXiv:2606. 19625v2 Announce Type: replace-cross Abstract: We use training-data attribution as an interpretable tool for capability discovery, mapping which regions of the pretraining corpus support social-reasoning versus STEM-reasoning in OLMo3-7B.
arXiv:2607. 14888v1 Announce Type: cross Abstract: Finetuning language models on small, curated datasets is standard practice for adapting them to specific policies or domains.