What and Whose Knowledge? Measuring Epistemic Diversity in Large Language Models
Measures the diversity of claims in language-model outputs across topics and cultural contexts.
- Communities, Domains & Consequences
- Factuality, Reliability & Evaluation
Publications
Publications
Recent group papers and selected earlier work by our faculty.
Includes work predating AAU appointments. For the institutional record, see AAU's research portal ↗.
82 publications
QQ: A Language Metadata Toolkit for Multilingual NLP
- Multilingual & Lower-Resource NLP
- Language, Meaning & Typology
TBLS: Transition-Based Language Similarity through Activation Perturbation
Measures language similarity through perturbation-induced cross-lingual transitions, finding stable structure across multilingual models and datasets.
- Multilingual & Lower-Resource NLP
- Language, Meaning & Typology
When Discourse Pressures Conflict: Information Structure in Vision-Language Model Outputs
Shows that vision-language models respond to discourse context in Hungarian but compress interacting information-structural pressures into narrow patterns.
- Language, Meaning & Typology
- Factuality, Reliability & Evaluation
Few-shot Semantic Recovery Attacks on Image Embeddings
Introduces a few-shot black-box method for recovering captions and sensitive attributes from image embeddings, and tests the utility cost of current defences.
- Security, Privacy & Safety
Do LLMs Really Memorize Personally Identifiable Information? Revisiting PII Leakage with a Cue-Controlled Memorization Framework
Separates genuine memorisation from reconstruction enabled by lexical cues across 32 languages, tightening the evidential standard for claims about PII leakage.
- Security, Privacy & Safety
- Multilingual & Lower-Resource NLP
MultiHal: Multilingual Dataset for Knowledge-Graph Grounded Evaluation of LLM Hallucinations
Builds a multilingual multihop benchmark from 25,900 curated knowledge-graph paths for evaluating and reducing hallucinations.
- Knowledge, Evidence & Reasoning
- Multilingual & Lower-Resource NLP
How Good is Your Wikipedia? Auditing Data Quality for Low-resource and Multilingual NLP
Audits all non-English Wikipedia editions and shows why data quality—not language coverage alone—must be a first-class concern in multilingual NLP.
- Multilingual & Lower-Resource NLP
- Factuality, Reliability & Evaluation
Stress Testing Factual Consistency Metrics for Long-Document Summarization
Demonstrates that commonly used factual-consistency metrics are brittle under controlled changes to long-document summaries.
- Knowledge, Evidence & Reasoning
- Factuality, Reliability & Evaluation
Follow the Path: Reasoning over Knowledge Graph Paths to Improve Large Language Model Factuality
Grounds reasoning traces in knowledge-graph paths, improving factual question answering particularly for smaller models and multihop questions.
- Knowledge, Evidence & Reasoning
- Factuality, Reliability & Evaluation
A Principled Framework for Evaluating on Typologically Diverse Languages
- Language, Meaning & Typology
- Multilingual & Lower-Resource NLP
Effective Performance Measurement: Challenges and Opportunities in KPI Extraction from Earnings Calls
- Knowledge, Evidence & Reasoning
- Communities, Domains & Consequences
Towards Understanding Professional AI Assistant Use: Human-in-the-Loop Topic Modeling of Multi-turn Conversations
- Factuality, Reliability & Evaluation
- Communities, Domains & Consequences
Real or Robotic? Assessing Whether LLMs Accurately Simulate Qualities of Human Responses in Human-LLM Dialogue
- Factuality, Reliability & Evaluation
- Communities, Domains & Consequences
NLP Security and Ethics, in the Wild
Audits ethical practice in NLP security and develops concrete guidance for harm minimisation and responsible disclosure.
- Security, Privacy & Safety
- Communities, Domains & Consequences
Shared Path: Unraveling Memorization in Multilingual LLMs through Language Similarities
- Security, Privacy & Safety
- Multilingual & Lower-Resource NLP
Large Language Models are Easily Confused: A Quantitative Metric, Security Implications and Typological Analysis
- Security, Privacy & Safety
- Multilingual & Lower-Resource NLP
InFact: Informativeness Alignment for Improved LLM Factuality
- Knowledge, Evidence & Reasoning
- Factuality, Reliability & Evaluation
Against All Odds: Overcoming Typology, Script, and Language Confusion in Multilingual Embedding Inversion Attacks
- Security, Privacy & Safety
- Multilingual & Lower-Resource NLP
ALGEN: Few-shot Inversion Attacks on Textual Embeddings via Cross-Model Alignment and Generation
- Security, Privacy & Safety
- Multilingual & Lower-Resource NLP
Limited-Resource Adapters Are Regularizers, Not Linguists
- Multilingual & Lower-Resource NLP
- Factuality, Reliability & Evaluation
Tokenization on Trial: The Case of Kalaallisut–Danish Legal Machine Translation
- Language, Meaning & Typology
- Multilingual & Lower-Resource NLP
A Cross-Lingual Perspective on Neural Machine Translation Difficulty
- Language, Meaning & Typology
- Multilingual & Lower-Resource NLP
Findings of the First Shared Task for Creole Language Machine Translation at WMT25
- Multilingual & Lower-Resource NLP
- Communities, Domains & Consequences
Linguistically Grounded Analysis of Language Models using Shapley Head Values
- Language, Meaning & Typology
- Factuality, Reliability & Evaluation
Knowledge Graphs, Large Language Models, and Hallucinations: An NLP Perspective
- Knowledge, Evidence & Reasoning
- Factuality, Reliability & Evaluation
Unstructured Evidence Attribution for Long Context Query Focused Summarization
- Knowledge, Evidence & Reasoning
- Factuality, Reliability & Evaluation
Efficiency Is Not Enough: A Critical Perspective on Environmentally Sustainable AI
- Communities, Domains & Consequences
Efficiency and Effectiveness of LLM-Based Summarization of Evidence in Crowdsourced Fact-Checking
- Knowledge, Evidence & Reasoning
- Communities, Domains & Consequences
Aggregating soft labels from crowd annotations improves uncertainty estimation under distribution shift
- Factuality, Reliability & Evaluation
- Communities, Domains & Consequences
On-Device LLMs for Home Assistant: Dual Role in Intent Detection and Response Generation
- Factuality, Reliability & Evaluation
- Communities, Domains & Consequences
How Do Hackathons Foster Creativity? Towards Automated Evaluation of Creativity at Scale
- Factuality, Reliability & Evaluation
- Communities, Domains & Consequences
The Responsible Development of Automated Student Feedback with Generative AI
- Factuality, Reliability & Evaluation
- Communities, Domains & Consequences
Utilizing phonetic similarity for cross-source and cross-language toponym matching: a benchmark and prototype
- Multilingual & Lower-Resource NLP
- Communities, Domains & Consequences
Putting AI Chatbots to the Test: A Comparative Analysis of Large Language Models’ Performance in the Context of Basal Cell Carcinoma
- Knowledge, Evidence & Reasoning
- Communities, Domains & Consequences
The Role of Typological Feature Prediction in NLP and Linguistics
- Language, Meaning & Typology
- Multilingual & Lower-Resource NLP
CreoleVal: Multilingual Multitask Benchmarks for Creoles
- Multilingual & Lower-Resource NLP
- Communities, Domains & Consequences
Text Embedding Inversion Security for Multilingual Language Models
- Security, Privacy & Safety
- Multilingual & Lower-Resource NLP
What is “Typological Diversity” in NLP?
- Multilingual & Lower-Resource NLP
- Factuality, Reliability & Evaluation
Multilingual Gradient Word-Order Typology from Universal Dependencies
- Language, Meaning & Typology
- Multilingual & Lower-Resource NLP
Sociolinguistically Informed Interpretability: A Case Study on Hinglish Emotion Classification
- Language, Meaning & Typology
- Multilingual & Lower-Resource NLP
A Call for Consistency in Reporting Typological Diversity
- Multilingual & Lower-Resource NLP
- Factuality, Reliability & Evaluation
Leveraging Adapters for Improved Cross-lingual Transfer for Low-Resource Creole MT
- Multilingual & Lower-Resource NLP
- Factuality, Reliability & Evaluation
LLMs Cannot (Yet) Match the Specificity and Simplicity of Online Communities in Long Form Question Answering
- Factuality, Reliability & Evaluation
- Communities, Domains & Consequences
Wiki-VEL: Visual Entity Linking for Structured Data on Wikimedia Commons
- Knowledge, Evidence & Reasoning
- Communities, Domains & Consequences
MADLINK: Attentive multihop and entity descriptions for link prediction in knowledge graphs
Combines graph paths with entity descriptions to predict missing links in knowledge graphs.
- Knowledge, Evidence & Reasoning
LLM Tropes: Revealing Fine-Grained Values and Opinions in Large Language Models
- Factuality, Reliability & Evaluation
- Communities, Domains & Consequences
Understanding Fine-grained Distortions in Reports of Scientific Findings
- Knowledge, Evidence & Reasoning
- Factuality, Reliability & Evaluation
BMRS: Bayesian Model Reduction for Structured Pruning
- Communities, Domains & Consequences
A chat about actinic keratosis: Examining capabilities and user experience of ChatGPT as a digital health technology in dermato-oncology
- Knowledge, Evidence & Reasoning
- Communities, Domains & Consequences
Recommending tasks based on search queries and missions
- Knowledge, Evidence & Reasoning
- Communities, Domains & Consequences
The Past, Present, and Future of Typological Databases in NLP
- Language, Meaning & Typology
- Multilingual & Lower-Resource NLP
Colex2Lang: Language Embeddings from Semantic Typology
- Language, Meaning & Typology
- Multilingual & Lower-Resource NLP
Do bridges dream of water pollutants? Towards DreamsKG, a knowledge graph to make digital access for sustainable environmental assessment come true
- Knowledge, Evidence & Reasoning
- Communities, Domains & Consequences
Multilingual Knowledge Graphs and Low-Resource Languages: A Review
- Knowledge, Evidence & Reasoning
- Multilingual & Lower-Resource NLP
Knowledge Graph Embeddings: Open Challenges and Opportunities
- Knowledge, Evidence & Reasoning
- Factuality, Reliability & Evaluation
Large Language Models and Knowledge Graphs: Opportunities and Challenges
- Knowledge, Evidence & Reasoning
- Factuality, Reliability & Evaluation
Quantifying Synthesis and Fusion and their Impact on Machine Translation
- Language, Meaning & Typology
- Multilingual & Lower-Resource NLP
What a Creole Wants, What a Creole Needs
- Multilingual & Lower-Resource NLP
- Communities, Domains & Consequences
Compositional Generalization in Multilingual Semantic Parsing over Wikidata
- Language, Meaning & Typology
- Multilingual & Lower-Resource NLP
Challenges and Strategies in Cross-Cultural NLP
- Multilingual & Lower-Resource NLP
- Communities, Domains & Consequences
Generating Scientific Claims for Zero-Shot Scientific Fact Checking
- Knowledge, Evidence & Reasoning
- Factuality, Reliability & Evaluation
Modeling Information Change in Science Communication with Semantically Matched Paraphrases
- Knowledge, Evidence & Reasoning
- Communities, Domains & Consequences
Entity Type Prediction Leveraging Graph Walks and Entity Descriptions
- Knowledge, Evidence & Reasoning
It’s All in the Name: Entity Typing Using Multilingual Language Models
- Knowledge, Evidence & Reasoning
- Multilingual & Lower-Resource NLP
Does Typological Blinding Impede Cross-Lingual Sharing?
- Language, Meaning & Typology
- Multilingual & Lower-Resource NLP
Inducing Language-Agnostic Multilingual Representations
- Multilingual & Lower-Resource NLP
- Factuality, Reliability & Evaluation
On Language Models for Creoles
- Multilingual & Lower-Resource NLP
- Language, Meaning & Typology
A survey on knowledge graph embeddings with literals: Which model links better literal-ly?
- Knowledge, Evidence & Reasoning
Cat2Type: Wikipedia Category Embeddings for Entity Typing in Knowledge Graphs
- Knowledge, Evidence & Reasoning
CiteWorth: Cite-Worthiness Detection for Improved Scientific Document Understanding
- Knowledge, Evidence & Reasoning
- Factuality, Reliability & Evaluation
Semi-Supervised Exaggeration Detection of Health Science Press Releases
- Knowledge, Evidence & Reasoning
- Factuality, Reliability & Evaluation
SubjQA: A Dataset for Subjectivity and Review Comprehension
- Knowledge, Evidence & Reasoning
- Factuality, Reliability & Evaluation
Zero-Shot Cross-Lingual Transfer with Meta Learning
- Multilingual & Lower-Resource NLP
- Factuality, Reliability & Evaluation
Back to the Future – Temporal Adaptation of Text Representations
- Language, Meaning & Typology
- Factuality, Reliability & Evaluation
What Do Language Representations Really Represent?
- Language, Meaning & Typology
- Multilingual & Lower-Resource NLP
Uncovering Probabilistic Implications in Typological Knowledge Bases
- Language, Meaning & Typology
- Knowledge, Evidence & Reasoning
A Probabilistic Generative Model of Linguistic Typology
- Language, Meaning & Typology
- Multilingual & Lower-Resource NLP
From Phonology to Syntax: Unsupervised Linguistic Typology at Different Levels with Language Embeddings
- Language, Meaning & Typology
- Multilingual & Lower-Resource NLP
The Parallel Meaning Bank: Towards a Multilingual Corpus of Translations Annotated with Compositional Meaning Representations
- Language, Meaning & Typology
- Multilingual & Lower-Resource NLP
Semantic Tagging with Deep Residual Networks
- Language, Meaning & Typology
- Multilingual & Lower-Resource NLP
The Meaning Factory: Formal Semantics for Recognizing Textual Entailment and Determining Semantic Similarity
- Language, Meaning & Typology
- Knowledge, Evidence & Reasoning
No matching publications
Try a broader search or clear one of the filters.
Scope
What is included here?
Selected peer-reviewed papers and accepted conference contributions, including earlier work by faculty before joining AAU. Author or research-group records are linked where proceedings are not yet available. Papers by other group members are included even when no faculty member is a coauthor.
Last updated September 2026. We omit duplicate preprint versions and do not claim complete coverage. Search by author name to find work by PhD fellows and research staff.