arXiv subjects

Orfeas Menis Mastromichalakis

Publications and source records attributed to Orfeas Menis Mastromichalakis.

16 recordsLinked to original sources

arXiv2026

Last Translation Benchmark

For scientific progress, we need benchmarks that test the limits of state-of-the-art models, and evaluation methods that inform us about failure cases. As models get stronger, standard benchmarks for machine translation are approaching saturation. Further, automatic translation metrics are unreliable, opaque, and vulnerable to reward-hacking. Even gold human evaluation is not problem-free, because it often lacks reproducibility, objectivity, and scalability. Overall, this prevents us from tracking progress in the field and identifying pathways for improvement. We introduce the Last Translation Benchmark, a collection of human-authored and peer-reviewed examples (texts, images, audio, videos) that break leading machine translation models. We also present a new evaluation approach: each example comes with handcrafted verification rules describing concrete failure cases on that example, therefore allowing reliable and actionable future evaluation. The Last Translation Benchmark is a live dataset that accepts ongoing contributions. The latest version is LTBv1, containing accepted contributions prior to September 1st 2026, with future releases planned as new data is continuously collected.

cs.CL↗

arXiv2026

Benchmarking Gender Bias in Machine Translation Evaluation Metrics across Occupations

Gender bias remains a persistent concern in machine translation (MT), affecting both generated translations and their automatic evaluation. When a source text leaves a person's gender unspecified, translations may realize that person using masculine or feminine forms, and both MT systems and evaluation metrics may exhibit systematic preferences between these alternatives despite the source providing no basis for such a distinction. We study this behavior in the WMT 2026 Automated Translation Quality Evaluation Systems Shared Task using an occupation-balanced subset of GAMBIT+. We consider seven English-source language pairs, six from the original dataset, targeting Arabic, Czech, Greek, Icelandic, Russian, and Ukrainian, and extend the original resource with German. The subset contains 1,308 masculine/feminine translation pairs per target language, with three examples for each of the 436 ISCO-08 occupational groups. We evaluate shared-task submissions and baselines for score prediction and error annotation, examining the direction, magnitude, and frequency of gender-related differences. We find an overall tendency for masculine translations to receive higher scores, as well as differences per occupation following stereotypical gender representations, although the strength and consistency of this preference vary considerably across evaluators and languages. Our results show that gender bias remains present in MT evaluation, but that capturing its extent requires looking beyond a single aggregate measure to complementary dimensions of evaluator behavior.

cs.CL↗

arXiv2026

An Empirical Study of Counterfactual Self-Explanations in LLMs

Large language models can easily generate explanations for their own outputs, but such self-explanations are not necessarily faithful to the model's behavior. We study this issue through counterfactual self-explanations, where a model minimally edits an input so that its own prediction changes. Across sentiment analysis and natural language inference, we evaluate ten instruction-tuned models from the LLaMA-3 and Qwen-2.5 families, measuring faithfulness, minimality, and alignment with human-annotated rationales. Our results show that model scale is the strongest determinant of explanation quality: larger models are substantially more likely to generate counterfactuals that flip their own predictions and target decision-relevant evidence. In contrast, the rationale-guided condition produces edit-minimal counterfactuals that are also more human-aligned. However, it does not consistently improve faithfulness. Overall, counterfactual self-explanations can provide useful behavioral evidence about model decisions, but their reliability depends strongly on model capacity and should be empirically validated rather than assumed.

cs.CL↗

arXiv2026

Harbor Adapters and Harbor-Index: Infrastructure and a Curated Meta-Dataset for Large-Scale Agentic Evaluation

Evaluating agents on the growing number of agentic benchmarks is challenging because they often require complex environments and agent integrations. We introduce Harbor Adapters, a unified evaluation infrastructure for agentic benchmarks. Our work makes three contributions. First, we develop benchmark adapters that port more than 80 benchmarks to evaluate arbitrary agents, and validate them through rigorous code review and parity experiments. Second, we conduct a large-scale evaluation of 8 models spanning capability tiers across 54 benchmarks; every model is run with Terminus-2 and with one of 3 native harnesses. This enables a broader analysis of agent capabilities and failure modes than was previously possible. Third, we introduce Harbor-Index, a curated set of 82 difficult, diverse, and high-quality tasks spanning 29 benchmarks, refined from the adapted suite through difficulty filtering, AI and human audit, and an audit-and-fix loop. Harbor-Index preserves the challenge and breadth of large-scale agentic evaluations while being affordable to run; no evaluated model-harness configuration exceeds 30% pass rate, and the strongest (GPT-5.5 with Codex) reaches 28.0%. We release the adapters, evaluation results, in-depth analysis, and Harbor-Index as open-source artifacts to support more reliable and comprehensive evaluation of language-model agents.

cs.AI↗

arXiv2026

AURA-Eval: Evaluation Framework for Acting Under Risk Awareness in LLM Agent Trajectories

LLM agents operate in workflows where unsafe actions can have real consequences. Existing safety evaluations often reduce behavior to a single score, obscuring risk recognition, pre-action detection, and safe task completion when a safe solution exists. We introduce AURA-Eval, a framework combining controlled augmentation with granular diagnosis of behavior in tool-use trajectories. Its pipeline identifies safety-critical decision points, generates controlled variations, and constructs counterparts differing in whether a request has a safe fulfillment path. Using 157 sourced trajectories, we generate 1,249 evaluation items and evaluate 20 frontier and open-weight models. We developed rubrics to classify risk detection, action strategy, and scenario-specific action safety. Our results show that LLM agents engage in unsafe behavior more often when no safe fulfillment path exists. In these cases, frontier proprietary models more often recognize risk and exhibit safer behavior by proposing alternatives, while evaluated open-weight models more often directly execute unsafe requests. Increasing impact or reducing opportunities for oversight before execution also exposes greater vulnerability across models.

cs.CR↗

arXiv2026

The Grounding Gap: How LLMs Anchor the Meaning of Abstract Concepts Differently from Humans

Abstract concepts - justice, theory, availability - have no single perceivable referent; in the human brain, their meaning emerges from a web of experiences, affect, and social context. Do large language models (LLMs) ground abstract concepts in a similar way? We study this by replicating property-generation experiments from cognitive science on 21 frontier and open-weight LLMs. Across models and experiments, we find a consistent pattern: when compared to humans, models rely too heavily on word associations, and underproduce properties tied to emotion and internal states. This yields a large and consistent grounding gap: no model exceeds a Pearson correlation r=0.37 with human responses, compared to a human-to-human ceiling above r=0.9. To better interpret this gap, we also replicate a rating experiment on grounding categories and find that here LLMs align more closely with human judgment, and alignment improves as models get larger. We then use sparse autoencoders (SAEs) to inspect whether this information is also reflected in the models' internal features, and we do identify features connected to grounding dimensions such as "sensorimotor" and "social". These findings suggest that current LLMs can recover grounding dimensions when explicitly queried, but do not recruit them in a human-like way when words are generated freely.

cs.CL↗

arXiv2026

Explain the Flag: Contextualizing Hate Speech Beyond Censorship

Hate, derogatory, and offensive speech remains a persistent challenge in online platforms and public discourse. While automated detection systems are widely used, most focus on censorship or removal, raising concerns for transparency and freedom of expression, and limiting opportunities to explain why content is harmful. To address these issues, explanatory approaches have emerged as a promising solution, aiming to make hate speech detection more transparent, accountable, and informative. In this paper, we present a hybrid approach that combines Large Language Models (LLMs) with three newly created and curated vocabularies to detect and explain hate speech in English, French, and Greek. Our system captures both inherently derogatory expressions tied to identity characteristics and direct group-targeted content through two complementary pipelines: one that detects and disambiguates problematic terms using the curated vocabularies, and one that leverages LLMs as context-aware evaluators of group-targeting content. The outputs are fused into grounded explanations that clarify why content is flagged. Human evaluation shows that our hybrid approach is accurate, with high-quality explanations, outperforming LLM-only baselines.

cs.CL↗

arXiv2026

Terminal-Bench: Benchmarking Agents on Hard, Realistic Tasks in Command Line Interfaces

AI agents may soon become capable of autonomously completing valuable, long-horizon tasks in diverse domains. Current benchmarks either do not measure real-world tasks, or are not sufficiently difficult to meaningfully measure frontier models. To this end, we present Terminal-Bench 2.0: a carefully curated hard benchmark composed of 89 tasks in computer terminal environments inspired by problems from real workflows. Each task features a unique environment, human-written solution, and comprehensive tests for verification. We show that frontier models and agents score less than 65\% on the benchmark and conduct an error analysis to identify areas for model and agent improvement. We publish the dataset and evaluation harness to assist developers and researchers in future work at https://www.tbench.ai/ .

cs.SE↗

arXiv2025

GAMBIT+: A Challenge Set for Evaluating Gender Bias in Machine Translation Quality Estimation Metrics

Gender bias in machine translation (MT) systems has been extensively documented, but bias in automatic quality estimation (QE) metrics remains comparatively underexplored. Existing studies suggest that QE metrics can also exhibit gender bias, yet most analyses are limited by small datasets, narrow occupational coverage, and restricted language variety. To address this gap, we introduce a large-scale challenge set specifically designed to probe the behavior of QE metrics when evaluating translations containing gender-ambiguous occupational terms. Building on the GAMBIT corpus of English texts with gender-ambiguous occupations, we extend coverage to three source languages that are genderless or natural-gendered, and eleven target languages with grammatical gender, resulting in 33 source-target language pairs. Each source text is paired with two target versions differing only in the grammatical gender of the occupational term(s) (masculine vs. feminine), with all dependent grammatical elements adjusted accordingly. An unbiased QE metric should assign equal or near-equal scores to both versions. The dataset's scale, breadth, and fully parallel design, where the same set of texts is aligned across all languages, enables fine-grained bias analysis by occupation and systematic comparisons across languages.

cs.CL↗

arXiv2025

Assumed Identities: Quantifying Gender Bias in Machine Translation of Gender-Ambiguous Occupational Terms

Machine Translation (MT) systems frequently encounter gender-ambiguous occupational terms, where they must assign gender without explicit contextual cues. While individual translations in such cases may not be inherently biased, systematic patterns-such as consistently translating certain professions with specific genders-can emerge, reflecting and perpetuating societal stereotypes. This ambiguity challenges traditional instance-level single-answer evaluation approaches, as no single gold standard translation exists. To address this, we introduce GRAPE, a probability-based metric designed to evaluate gender bias by analyzing aggregated model responses. Alongside this, we present GAMBIT, a benchmarking dataset in English with gender-ambiguous occupational terms. Using GRAPE, we evaluate several MT systems and examine whether their gendered translations in Greek and French align with or diverge from societal stereotypes, real-world occupational gender distributions, and normative standards

cs.CL↗

arXiv2025

Don't Erase, Inform! Detecting and Contextualizing Harmful Language in Cultural Heritage Collections

Cultural Heritage (CH) data hold invaluable knowledge, reflecting the history, traditions, and identities of societies, and shaping our understanding of the past and present. However, many CH collections contain outdated or offensive descriptions that reflect historical biases. CH Institutions (CHIs) face significant challenges in curating these data due to the vast scale and complexity of the task. To address this, we develop an AI-powered tool that detects offensive terms in CH metadata and provides contextual insights into their historical background and contemporary perception. We leverage a multilingual vocabulary co-created with marginalized communities, researchers, and CH professionals, along with traditional NLP techniques and Large Language Models (LLMs). Available as a standalone web app and integrated with major CH platforms, the tool has processed over 7.9 million records, contextualizing the contentious terms detected in their metadata. Rather than erasing these terms, our approach seeks to inform, making biases visible and providing actionable insights for creating more inclusive and accessible CH collections.

cs.CL↗

arXiv2025

MusicLIME: Explainable Multimodal Music Understanding

Multimodal models are critical for music understanding tasks, as they capture the complex interplay between audio and lyrics. However, as these models become more prevalent, the need for explainability grows-understanding how these systems make decisions is vital for ensuring fairness, reducing bias, and fostering trust. In this paper, we introduce MusicLIME, a model-agnostic feature importance explanation method designed for multimodal music models. Unlike traditional unimodal methods, which analyze each modality separately without considering the interaction between them, often leading to incomplete or misleading explanations, MusicLIME reveals how audio and lyrical features interact and contribute to predictions, providing a holistic view of the model's decision-making. Additionally, we enhance local explanations by aggregating them into global explanations, giving users a broader perspective of model behavior. Through this work, we contribute to improving the interpretability of multimodal music models, empowering users to make informed choices, and fostering more equitable, fair, and transparent music understanding systems.

cs.SD↗

arXiv2025

AILS-NTUA at SemEval-2025 Task 4: Parameter-Efficient Unlearning for Large Language Models using Data Chunking

The Unlearning Sensitive Content from Large Language Models task aims to remove targeted datapoints from trained models while minimally affecting their general knowledge. In our work, we leverage parameter-efficient, gradient-based unlearning using low-rank (LoRA) adaptation and layer-focused fine-tuning. To further enhance unlearning effectiveness, we employ data chunking, splitting forget data into disjoint partitions and merging them with cyclically sampled retain samples at a pre-defined ratio. Our task-agnostic method achieves an outstanding forget-retain balance, ranking first on leaderboards and significantly outperforming baselines and competing systems.

cs.CL↗

arXiv2024

GOSt-MT: A Knowledge Graph for Occupation-related Gender Biases in Machine Translation

Gender bias in machine translation (MT) systems poses significant challenges that often result in the reinforcement of harmful stereotypes. Especially in the labour domain where frequently occupations are inaccurately associated with specific genders, such biases perpetuate traditional gender stereotypes with a significant impact on society. Addressing these issues is crucial for ensuring equitable and accurate MT systems. This paper introduces a novel approach to studying occupation-related gender bias through the creation of the GOSt-MT (Gender and Occupation Statistics for Machine Translation) Knowledge Graph. GOSt-MT integrates comprehensive gender statistics from real-world labour data and textual corpora used in MT training. This Knowledge Graph allows for a detailed analysis of gender bias across English, French, and Greek, facilitating the identification of persistent stereotypes and areas requiring intervention. By providing a structured framework for understanding how occupations are gendered in both labour markets and MT systems, GOSt-MT contributes to efforts aimed at making MT systems more equitable and reducing gender biases in automated translations.

cs.CL↗

arXiv2024

Beyond One-Size-Fits-All: Adapting Counterfactual Explanations to User Objectives

Explainable Artificial Intelligence (XAI) has emerged as a critical area of research aimed at enhancing the transparency and interpretability of AI systems. Counterfactual Explanations (CFEs) offer valuable insights into the decision-making processes of machine learning algorithms by exploring alternative scenarios where certain factors differ. Despite the growing popularity of CFEs in the XAI community, existing literature often overlooks the diverse needs and objectives of users across different applications and domains, leading to a lack of tailored explanations that adequately address the different use cases. In this paper, we advocate for a nuanced understanding of CFEs, recognizing the variability in desired properties based on user objectives and target applications. We identify three primary user objectives and explore the desired characteristics of CFEs in each case. By addressing these differences, we aim to design more effective and tailored explanations that meet the specific needs of users, thereby enhancing collaboration with AI systems.

cs.LG↗

arXiv2022

Towards explainable evaluation of language models on the semantic similarity of visual concepts

Recent breakthroughs in NLP research, such as the advent of Transformer models have indisputably contributed to major advancements in several tasks. However, few works research robustness and explainability issues of their evaluation strategies. In this work, we examine the behavior of high-performing pre-trained language models, focusing on the task of semantic similarity for visual vocabularies. First, we address the need for explainable evaluation metrics, necessary for understanding the conceptual quality of retrieved instances. Our proposed metrics provide valuable insights in local and global level, showcasing the inabilities of widely used approaches. Secondly, adversarial interventions on salient query semantics expose vulnerabilities of opaque metrics and highlight patterns in learned linguistic representations.

cs.CL↗