Rita Sevastjanova

Hi! I am Rita Sevastjanova.

I am a data analyst with a visual analytics background, specializing in human-centered explainability for large language models. My work combines computational methods with interactive visual interfaces to bridge the gap between complex model internals and human understanding.

I earned my Ph.D. in Natural Sciences (Computer and Information Science) at the University of Konstanz in 2023.

Projects

PhD Research

During my PhD, I developed visual and interactive approaches to investigate what information is learned by large language models, including linguistic properties (e.g., semantics and syntax) and undesired biases (e.g., gender bias), with several approaches tailored to model comparison.

Undergraduate & Graduate Research

During my bachelor and master studies, I worked as a research assistant, leading the development of the lingvis.io framework for rapid prototyping of linguistic, web-based visual analytics applications. Both my bachelor's and master's theses were conducted as part of this framework.

lingvis.io

Visual Analytics. Linguistics. Machine Learning.

Events

I have given multiple invited talks and taken part in panel discussions focused on large language models, discussing their use in industry applications alongside critical limitations such as bias and other risks.

King's College London

Invited talk

Visual Exploration of (Large) Language Models

March 2025, Online

Collegium Helveticum

Panel discussion and invited talk

Visual Analytics Approaches for Exploring Large Language Models

January 2025, Zürich, Switzerland

IPAI Experience Day

Panel and invited talk

Investigating Biases in AI through Interactive Machine Learning

October 2024, Heilbronn, Germany

Digitalgipfel der Bundesregierung

Panel discussion

KI im Tagesgeschäft? – Erfolgsrezepte aus Unternehmen, öffentlicher Verwaltung und Politik

October 2024, Frankfurt am Main, Germany

TED AI

Workshop

Discovering the Power of Human-AI Collaboration Firsthand

October 2024, Vienna, Austria

Google Cloud Partner Connect

Invited talk

Bias in Generative Language Models

September 2024, Zürich, Switzerland

Collective Gathering at the Startup Days

Panel discussion

Empathy in AI: From Tool to Human Friend

May 2024, Bern, Switzerland

Publications

All Publications

Revealing the Unwritten: Visual Investigation of Beam Search Trees to Address Language Model Prompting Challenges

T. Spinner, R. Kehlbeck, R. Sevastjanova, T. Stähle, D.A. Keim, O. Deussen, M. El-Assady
Proc. of the Association for Computational Linguistics (System Demonstrations) 2025
bias explainability nlp visualization

Finding Needles in Document Haystacks: Augmenting Serendipitous Claim Retrieval Workflows

M. Dück, S. Holter, R.S.M. Chan, R. Sevastjanova, M. El-Assady
Proc. of the 2025 CHI Conference on Human Factors in Computing Systems 2025
text analysis visualization

Concept-Level Explainability for Auditing & Steering LLM Responses

K. Amara, R. Sevastjanova, M. El-Assady
arXiv preprint arXiv:2505.07610 2025
bias explainability nlp

A multi-cue study to the interpretation of German information-seeking and rhetorical questions

B. Braun, N. Dehé, M. Einfeldt, A. James, E. Kazak, R. Sevastjanova, K. Zahner-Ritter, E. Kazak
Laboratory Phonology 16 (1) 2025
active learning nlp

LayerFlow: Layer-wise Exploration of LLM Embeddings using Uncertainty-aware Interlinked Projections

R. Sevastjanova, R. Gerling, T. Spinner, M. El-Assady
Computer Graphics Forum 2025
explainability nlp uncertainty visualization

Explainable Mapper: Charting LLM Embedding Spaces Using Perturbation-Based Explanation and Verification Agents

X. Yan, R. Sevastjanova, S. van der Ben, M. El-Assady, B. Wang
arXiv preprint arXiv:2507.18607 2025
AI agents explainability nlp visualization

Is General-Purpose AI Reasoning Sensitive to Data-Induced Cognitive Biases? Dynamic Benchmarking on Typical Software Engineering Dilemmas

F. Sovrano, G. Dominici, R. Sevastjanova, A. Stramiglio, A. Bacchelli
arXiv preprint arXiv:2508.11278 2025
bias explainability nlp

Dia-Lingle: A Gamified Interface for Dialectal Data Collection

J. Sun, R. Sevastjanova, S. Ahmadi, R. Sennrich, M. El-Assady
Proc. of the Association for Computational Linguistics (System Demonstrations) 2025
active learning gamification nlp

A Design Space for Intelligent Agents in Mixed-Initiative Visual Analytics

T. Stähle, S. Boyer, R. Sevastjanova, A. Narechania, M. El-Assady
arXiv preprint arXiv:2512.23372 2025
ai agents explainability visualization

Personalized Language Model Selection Through Gamified Elicitation of Contrastive Concept Preferences

R. Sevastjanova, H. Hauptmann, S. Deterding, M. El-Assady
IEEE Trans. on Visualization and Computer Graphics 2024
gamification nlp visualization

The use of Active Learning systems for stimulus selection and response modelling in perception experiments

M. Einfeldt, R. Sevastjanova, K. Zahner-Ritter, E. Kazak, B. Braun
Computer Speech & Language 2024
active learning nlp

Syntaxshap: Syntax-aware explainability method for text generation

K. Amara, R. Sevastjanova, M. El-Assady
Findings of the Association for Computational Linguistics 2024
explainability nlp

Treeducation: A visual education platform for teaching treemap layout algorithms

J. Fuchs, B. Jäckl, M. Jüttler, D.A. Keim, R. Sevastjanova
IEEE Transactions on Visualization and Computer Graphics 2024
gamification visualization

Challenges and opportunities in text generation explainability

K. Amara, R. Sevastjanova, M. El-Assady
World Conference on Explainable Artificial Intelligence 2024
explainability nlp text analysis

Feature Clock: High-Dimensional Effects in Two-Dimensional Plots

O. Ovcharenko, R. Sevastjanova, V. Boeva
2024 IEEE Visualization and Visual Analytics (VIS) 2024
uncertainty visualization

WEC-Explainer: A Descriptive Framework for Exploring Word Embedding Contextualization

R. Sevastjanova, M. El-Assady
Exploring Research Opportunities for Natural Language, Text, and Data Visualization (NLVIZ) Workshop at IEEE VIS 2023
explainability nlp visualization

Visual Comparison of Text Sequences Generated by Large Language Models

R. Sevastjanova, S. Vogelbacher, A. Spitz, D. Keim, M. El-Assady
2023 IEEE Visualization in Data Science (VDS) 2023
bias nlp visualization

The Performance of International Organizations: A New Measure and Dataset Based on Computational Text Analysis of Evaluation Reports

S. Eckhard, V. Jankauskas, E. Leuschner, I. Burton, T. Kerl, R. Sevastjanova
The Review of International Organizations 2023
text analysis

MediCoSpace: Visual Decision-Support for Doctor-Patient Consultations Using Medical Concept Spaces from EHRs

S. van der Linden, R. Sevastjanova, M. Funk, M. El-Assady
ACM Trans. on Management Information Systems 2023
text analysis visualization

Visual Comparison of Language Model Adaptation

R. Sevastjanova, E. Cakmak, S. Ravfogel, R. Cotterell, M. El-Assady
IEEE Trans. on Visualization and Computer Graphics 2022
bias explainability nlp visualization

Beware the Rationalization Trap! When Language Model Explainability Diverges from our Mental Models of Language

R. Sevastjanova, M. El-Assady
Communication in Human-AI Interaction Workshop at IJCAI-ECAI'22 2022
bias explainability nlp

LMFingerprints: Visual Explanations of Language Model Embedding Spaces through Layerwise Contextualization Scores

R. Sevastjanova, A.-L. Kalouli, C. Beck, H. Hauptmann, M. El-Assady
Computer Graphics Forum 2022
explainability nlp visualization

Visualisierung der Covid-19-Inzidenzen und Behandlungskapazitäten mit Coronavis

W. Jentner, F. Sperrle, D. Seebacher, M. Kraus, R. Sevastjanova, et al.
Resilienz und Pandemie: Handlungsempfehlungen anhand von Erfahrungen mit COVID-19 2022
visualization

VisInReport: Complementing Visual Discourse Analytics Through Personalized Insight Reports

R. Sevastjanova, M. El-Assady, A. Bradley, C. Collins, M. Butt, D. Keim
IEEE Trans. on Visualization and Computer Graphics 2022
text analysis visualization

Explaining Contextualization in Language Models using Visual Analytics

R. Sevastjanova, A.-L. Kalouli, C. Beck, H. Hauptmann, M. El-Assady
Proc. of the Association for Computational Linguistics 2021
explainability nlp visualization

CommAID: Visual Analytics for Communication Analysis through Interactive Dynamics Modeling

M.T. Fischer, D. Seebacher, R. Sevastjanova, D.A. Keim, M. El-Assady
Computer Graphics Forum 2021
text analysis visualization

Reliable Estimates of Interpretable Cue Effects with Active Learning in Psycholinguistic Research

M. Einfeldt, R. Sevastjanova, K. Zahner-Ritter, E. Kazak, B. Braun
Interspeech 2021 2021
active learning nlp

Is that Really a Question?: Going beyond Factoid Questions in NLP

A.-L. Kalouli, R. Kehlbeck, R. Sevastjanova, O. Deussen, D.A. Keim, M. Butt
14th Int. Conf. on Computational Semantics: IWCS 2021 2021
nlp

QuestionComb: A Gamification Approach for the Visual Explanation of Linguistic Phenomena through Interactive Labeling

R. Sevastjanova, W. Jentner, F. Sperrle, R. Kehlbeck, J. Bernard, M. El-Assady
ACM Trans. on Interactive Intelligent Systems 2021
active learning gamification nlp visualization

Shall we play? – Extending the Visual Analytics Design Space through Gameful Design Concepts

R. Sevastjanova, H. Schäfer, J. Bernard, D. Keim, M. El-Assady
MLUI 2019: Machine Learning from User Interactions for Visualization and Analytics, IEEE VIS Workshop 2019
gamification visualization

Visual Analytics of Conversational Dynamics

D. Seebacher, M.T. Fischer, R. Sevastjanova, D.A. Keim, M. El-Assady
EuroVis Workshop on Visual Analytics (EuroVA) 2019
text analysis visualization

VIANA: Visual Interactive Annotation of Argumentation

F. Sperrle, R. Sevastjanova, R. Kehlbeck, M. El-Assady
2019 IEEE Conf. on Visual Analytics Science and Technology (VAST) 2019
text analysis visualization

ParHistVis: Visualization of Parallel Multilingual Historical Data

A.-L. Kalouli, R. Kehlbeck, R. Sevastjanova, K. Kaiser, G.A. Kaiser, M. Butt
Proc. of the 1st Int. Workshop on Computational Approaches to Historical Language Change 2019
nlp text analysis visualization

NEAT-Novel Emergency Analysis Tool

W. Jentner, J. Buchmüller, F. Sperrle, R. Sevastjanova, T. Spinner, U. Schlegel, D. Streeb, H. Schäfer
2019 IEEE Conf. on Visual Analytics Science and Technology (VAST) 2019
visualization

ThreadReconstructor: Modeling Reply-Chains to Untangle Conversational Text through Visual Analytics

M. El-Assady, R. Sevastjanova, D. Keim, C. Collins
Computer Graphics Forum 2018
text analysis visualization

LTMA: Layered Topic Matching for the Comparative Exploration, Evaluation, and Refinement of Topic Modeling Results

M. El-Assady, F. Sperrle, R. Sevastjanova, M. Sedlmair, D. Keim
2018 Int. Symposium on Big Data Visual and Immersive Analytics (BDVA) 2018
text analysis visualization

Going beyond Visualization: Verbalization as Complementary Medium to Explain Machine Learning Models

R. Sevastjanova, F. Beck, B. Ell, C. Turkay, R. Henkin, M. Butt, D.A. Keim, M. El-Assady
Workshop on Visualization for AI Explainability at IEEE VIS 2018
explainability visualization

Mixed-initiative Active Learning for Generating Linguistic Insights in Question Classification

R. Sevastjanova, M. El-Assady, A. Hautli-Janisz, A.-L. Kalouli, R. Kehlbeck, O. Deussen, D.A. Keim, M. Butt
Workshop on Data Systems for Interactive Analysis (DSIA) at IEEE VIS 2018
active learning nlp

Minions, Sheep, and Fruits: Metaphorical Narratives to Explain Artificial Intelligence and Build Trust

W. Jentner, R. Sevastjanova, F. Stoffel, D.A. Keim, J. Bernard, M. El-Assady
Workshop on Visualization for AI Explainability at IEEE 2018
explainability visualization

NEREx: Named-Entity Relationship Exploration in Multi-Party Conversations

M. El-Assady, R. Sevastjanova, B. Gipp, D. Keim, C. Collins
Computer Graphics Forum 2017
nlp text analysis visualization

Progressive Learning of Topic Modeling Parameters: A Visual Analytics Framework

M. El-Assady, R. Sevastjanova, F. Sperrle, D. Keim, C. Collins
IEEE Trans. on Visualization and Computer Graphics 2017
text analysis visualization