Jabez Magomere

William Lugoloobi, Samuelle Marro, Jabez Magomere et al.

As LLM-based agents increasingly browse the web on users' behalf, a natural question arises: can websites passively identify which underlying model powers an agent? Doing so would represent a significant security risk, enabling targeted attacks tailored to known model vulnerabilities. Across 14 frontier LLMs and four web environments spanning information retrieval and shopping tasks, we show that an agent's actions and interaction timings, captured via a passive JavaScript tracker, are sufficient to identify the underlying model with up to 96\% F1. We formalise this attack surface by demonstrating that classifiers trained on agent actions generalise across model sizes and families. We further show that strong classifiers can be trained from few interaction traces and that agent identity can be inferred early within an episode. Injecting randomised timing delays between actions substantially degrades classifier performance, but does not provide robust protection: a classifier retrained on delayed traces largely recovers performance. We release our harness and a labelled corpus of agent traces \href{https://github.com/KabakaWilliam/known_actions}{here}.

Andrew M. Bean, Ryan Othniel Kearns, Angelika Romanou et al.

Evaluating large language models (LLMs) is crucial for both assessing their capabilities and identifying safety or robustness issues prior to deployment. Reliably measuring abstract and complex phenomena such as 'safety' and 'robustness' requires strong construct validity, that is, having measures that represent what matters to the phenomenon. With a team of 29 expert reviewers, we conduct a systematic review of 445 LLM benchmarks from leading conferences in natural language processing and machine learning. Across the reviewed articles, we find patterns related to the measured phenomena, tasks, and scoring metrics which undermine the validity of the resulting claims. To address these shortcomings, we provide eight key recommendations and detailed actionable guidance to researchers and practitioners in developing LLM benchmarks.

Jabez Magomere, Elena Kochkina, Samuel Mensah et al.

Complex claim verification requires decomposing sentences into verifiable subclaims, yet existing methods struggle to align decomposition quality with verification performance. We propose a reinforcement learning (RL) approach that jointly optimizes decomposition quality and verifier alignment using Group Relative Policy Optimization (GRPO). Our method integrates: (i) structured sequential reasoning; (ii) supervised finetuning on teacher-distilled exemplars; and (iii) a multi-objective reward balancing format compliance, verifier alignment, and decomposition quality. Across six evaluation settings, our trained 8B decomposer improves downstream verification performance to (71.75%) macro-F1, outperforming prompt-based approaches ((+1.99), (+6.24)) and existing RL methods ((+5.84)). Human evaluation confirms the high quality of the generated subclaims. Our framework enables smaller language models to achieve state-of-the-art claim verification by jointly optimising for verification accuracy and decomposition quality.

Jabez Magomere, Emanuele La Malfa, Manuel Tonneau et al. · oxford

Online misinformation remains a critical challenge, and fact-checkers increasingly rely on claim matching systems that use sentence embedding models to retrieve relevant fact-checks. However, as users interact with claims online, they often introduce edits, and it remains unclear whether current embedding models used in retrieval are robust to such edits. To investigate this, we introduce a perturbation framework that generates valid and natural claim variations, enabling us to assess the robustness of a wide-range of sentence embedding models in a multi-stage retrieval pipeline and evaluate the effectiveness of various mitigation approaches. Our evaluation reveals that standard embedding models exhibit notable performance drops on edited claims, while LLM-distilled embedding models offer improved robustness at a higher computational cost. Although a strong reranker helps to reduce the performance drop, it cannot fully compensate for first-stage retrieval gaps. To address these retrieval gaps, we evaluate train- and inference-time mitigation approaches, demonstrating that they can improve in-domain robustness by up to 17 percentage points and boost out-of-domain generalization by 10 percentage points. Overall, our findings provide practical improvements to claim-matching systems, enabling more reliable fact-checking of evolving misinformation. Code and data are available at https://github.com/JabezNzomo99/claim-matching-robustness.

Jabez Magomere, Shu Ishida, Tejumade Afonja et al.

We introduce the World Wide recipe, which sets forth a framework for culturally aware and participatory data collection, and the resultant regionally diverse World Wide Dishes evaluation dataset. We also analyse bias operationalisation to highlight how current systems underperform across several dimensions: (in-)accuracy, (mis-)representation, and cultural (in-)sensitivity, with evidence from qualitative community-based observations and quantitative automated tools. We find that these T2I models generally do not produce quality outputs of dishes specific to various regions. This is true even for the US, which is typically considered more well-resourced in training data -- although the generation of US dishes does outperform that of the investigated African countries. The models demonstrate the propensity to produce inaccurate and culturally misrepresentative, flattening, and insensitive outputs. These representational biases have the potential to further reinforce stereotypes and disproportionately contribute to erasure based on region. The dataset and code are available at https://github.com/oxai/world-wide-dishes.

Jabez Magomere, Elena Kochkina, Samuel Mensah et al.

We introduce FinNLI, a benchmark dataset for Financial Natural Language Inference (FinNLI) across diverse financial texts like SEC Filings, Annual Reports, and Earnings Call transcripts. Our dataset framework ensures diverse premise-hypothesis pairs while minimizing spurious correlations. FinNLI comprises 21,304 pairs, including a high-quality test set of 3,304 instances annotated by finance experts. Evaluations show that domain shift significantly degrades general-domain NLI performance. The highest Macro F1 scores for pre-trained (PLMs) and large language models (LLMs) baselines are 74.57% and 78.62%, respectively, highlighting the dataset's difficulty. Surprisingly, instruction-tuned financial LLMs perform poorly, suggesting limited generalizability. FinNLI exposes weaknesses in current LLMs for financial reasoning, indicating room for improvement.

Samuel Mensah, Elena Kochkina, Jabez Magomere et al.

Mitigating entity bias is a critical challenge in Relation Extraction (RE), where models often rely excessively on entities, resulting in poor generalization. This paper presents a novel approach to address this issue by adapting a Variational Information Bottleneck (VIB) framework. Our method compresses entity-specific information while preserving task-relevant features. It achieves state-of-the-art performance on relation extraction datasets across general, financial, and biomedical domains, in both indomain (original test sets) and out-of-domain (modified test sets with type-constrained entity replacements) settings. Our approach offers a robust, interpretable, and theoretically grounded methodology.

Jabez Magomere, Scott Hale

The adoption of crowdsourced election monitoring as a complementary alternative to traditional election monitoring is on the rise. Yet, its reliance on digital response volunteers to manually process incoming election reports poses a significant scaling bottleneck. In this paper, we address the challenge of scaling crowdsourced election monitoring by advancing the task of automated classification of crowdsourced election reports to multilingual and cross-domain classification settings. We propose a two-step classification approach of first identifying informative reports and then categorising them into distinct information types. We conduct classification experiments using multilingual transformer models such as XLM-RoBERTa and multilingual embeddings such as SBERT, augmented with linguistically motivated features. Our approach achieves F1-Scores of 77\% for informativeness detection and 75\% for information type classification. We conduct cross-domain experiments, applying models trained in a source electoral domain to a new target electoral domain in zero-shot and few-shot classification settings. Our results show promising potential for model transfer across electoral domains, with F1-Scores of 59\% in zero-shot and 63\% in few-shot settings. However, our analysis also reveals a performance bias in detecting informative English reports over Swahili, likely due to imbalances in the training data, indicating a need for caution when deploying classification models in real-world election scenarios.

Andrew M. Bean, Simi Hellsten, Harry Mayne et al.

In this paper, we present the LingOly benchmark, a novel benchmark for advanced reasoning abilities in large language models. Using challenging Linguistic Olympiad puzzles, we evaluate (i) capabilities for in-context identification and generalisation of linguistic patterns in very low-resource or extinct languages, and (ii) abilities to follow complex task instructions. The LingOly benchmark covers more than 90 mostly low-resource languages, minimising issues of data contamination, and contains 1,133 problems across 6 formats and 5 levels of human difficulty. We assess performance with both direct accuracy and comparison to a no-context baseline to penalise memorisation. Scores from 11 state-of-the-art LLMs demonstrate the benchmark to be challenging, and models perform poorly on the higher difficulty problems. On harder problems, even the top model only achieved 38.7% accuracy, a 24.7% improvement over the no-context baseline. Large closed models typically outperform open models, and in general, the higher resource the language, the better the scores. These results indicate, in absence of memorisation, true multi-step out-of-domain reasoning remains a challenge for current language models.