arXiv subjects

Tobias Hecking

Publications and source records attributed to Tobias Hecking.

8 recordsLinked to original sources

arXiv2026

The Effectiveness of Style Vectors for Steering Large Language Models: A Human Evaluation

Controlling the behavior of large language models (LLMs) at inference time is essential for aligning outputs with human abilities and safety requirements. \emph{Activation steering} provides a lightweight alternative to prompt engineering and fine-tuning by directly modifying internal activations to guide generation. This research advances the literature in three significant directions. First, while previous work demonstrated the technical feasibility of steering emotional tone using automated classifiers, this paper presents the first human evaluation of activation steering concerning the emotional tone of LLM outputs, collecting over 7,000 crowd-sourced ratings from 190 participants via Prolific ($n=190$). These ratings assess both perceived emotional intensity and overall text quality. Second, we find strong alignment between human and model-based quality ratings (mean $r=0.776$, range $0.157$--$0.985$), indicating automatic scoring can proxy perceived quality. Moderate steering strengths ($\lambda \approx 0.15$) reliably amplify target emotions while preserving comprehensibility, with the strongest effects for disgust ($\eta_p^2 = 0.616$) and fear ($\eta_p^2 = 0.540$), and minimal effects for surprise ($\eta_p^2 = 0.042$). Finally, upgrading from Alpaca to LlaMA-3 yielded more consistent steering with significant effects across emotions and strengths (all $p < 0.001$). Inter-rater reliability was high (ICC $= 0.71$--$0.87$), underscoring the robustness of the findings. These findings support activation-based control as a scalable method for steering LLM behavior across affective dimensions.

cs.AI↗

arXiv2025

Integrating Knowledge Graphs and Visualization Dashboards for Advance Data Discovery in VESA

The increasing complexity and scale of scientific datasets demand advanced tools for efficient discovery and exploration. Traditional search systems often fall short in addressing the multidimensional nature of data and their intricate relationships, limiting their utility for researchers. This paper introduces the Knowledge Graph Based Visualization Search Application (VESA), which reshapes the process of data discovery by leveraging knowledge graph technology to establish meaningful connections and employing a visualization dashboard to enable multidimensional exploration. A software prototype is developed, showcasing our use case of connecting two Earth System Science repositories via a knowledge graph backend and visualization dashboard at the frontend. The framework's effectiveness was assessed against guidelines derived from a comprehensive literature review and further validated through an online user study. The evaluation revealed positive reception, highlighting VESA's low learning curve, ease of use, and potential to enhance data discovery workflows.

cs.DB↗

arXiv2024

Towards a Reliable Offline Personal AI Assistant for Long Duration Spaceflight

As humanity prepares for new missions to the Moon and Mars, astronauts will need to operate with greater autonomy, given the communication delays that make real-time support from Earth difficult. For instance, messages between Mars and Earth can take up to 24 minutes, making quick responses impossible. This limitation poses a challenge for astronauts who must rely on in-situ tools to access the large volume of data from spacecraft sensors, rovers, and satellites, data that is often fragmented and difficult to use. To bridge this gap, systems like the Mars Exploration Telemetry-Driven Information System (METIS) are being developed. METIS is an AI assistant designed to handle routine tasks, monitor spacecraft systems, and detect anomalies, all while reducing the reliance on mission control. Current Generative Pretrained Transformer (GPT) Models, while powerful, struggle in safety-critical environments. They can generate plausible but incorrect responses, a phenomenon known as "hallucination," which could endanger astronauts. To overcome these limitations, this paper proposes enhancing systems like METIS by integrating GPTs, Retrieval-Augmented Generation (RAG), Knowledge Graphs (KGs), and Augmented Reality (AR). The idea is to allow astronauts to interact with their data more intuitively, using natural language queries and visualizing real-time information through AR. KGs will be used to easily access live telemetry and multimodal data, ensuring that astronauts have the right information at the right time. By combining AI, KGs, and AR, this new system will empower astronauts to work more autonomously, safely, and efficiently during future space missions.

cs.AI↗

arXiv2024

AI Assistants for Spaceflight Procedures: Combining Generative Pre-Trained Transformer and Retrieval-Augmented Generation on Knowledge Graphs With Augmented Reality Cues

This paper describes the capabilities and potential of the intelligent personal assistant (IPA) CORE (Checklist Organizer for Research and Exploration), designed to support astronauts during procedures onboard the International Space Station (ISS), the Lunar Gateway station, and beyond. We reflect on the importance of a reliable and flexible assistant capable of offline operation and highlight the usefulness of audiovisual interaction using augmented reality elements to intuitively display checklist information. We argue that current approaches to the design of IPAs in space operations fall short of meeting these criteria. Therefore, we propose CORE as an assistant that combines Knowledge Graphs (KGs), Retrieval-Augmented Generation (RAG) for a Generative Pre-Trained Transformer (GPT), and Augmented Reality (AR) elements to ensure an intuitive understanding of procedure steps, reliability, offline availability, and flexibility in terms of response style and procedure updates.

cs.AI↗

arXiv2024

Style Vectors for Steering Generative Large Language Model

This research explores strategies for steering the output of large language models (LLMs) towards specific styles, such as sentiment, emotion, or writing style, by adding style vectors to the activations of hidden layers during text generation. We show that style vectors can be simply computed from recorded layer activations for input texts in a specific style in contrast to more complex training-based approaches. Through a series of experiments, we demonstrate the effectiveness of activation engineering using such style vectors to influence the style of generated text in a nuanced and parameterisable way, distinguishing it from prompt engineering. The presented research constitutes a significant step towards developing more adaptive and effective AI-empowered interactive systems.

cs.CL↗

arXiv2023

Predicting Winning Regions in Parity Games via Graph Neural Networks (Extended Abstract)

Solving parity games is a major building block for numerous applications in reactive program verification and synthesis. While they can be solved efficiently in practice, no known approach has a polynomial worst-case runtime complexity. We present a incomplete polynomial-time approach to determining the winning regions of parity games via graph neural networks. Our evaluation on 900 randomly generated parity games shows that this approach is effective and efficient in practice. It correctly determines the winning regions of $\sim$60\% of the games in our data set and only incurs minor errors in the remaining ones. We believe that this approach can be extended to efficiently solve parity games as well.

cs.GT↗

arXiv2021

Community Detection for Access-Control Decisions: Analysing the Role of Homophily and Information Diffusion in Online Social Networks

Access-Control Lists (ACLs) (a.k.a. friend lists) are one of the most important privacy features of Online Social Networks (OSNs) as they allow users to restrict the audience of their publications. Nevertheless, creating and maintaining custom ACLs can introduce a high cognitive burden on average OSNs users since it normally requires assessing the trustworthiness of a large number of contacts. In principle, community detection algorithms can be leveraged to support the generation of ACLs by mapping a set of examples (i.e. contacts labelled as untrusted) to the emerging communities inside the user's ego-network. However, unlike users' access-control preferences, traditional community-detection algorithms do not take the homophily characteristics of such communities into account (i.e. attributes shared among members). Consequently, this strategy may lead to inaccurate ACL configurations and privacy breaches under certain homophily scenarios. This work investigates the use of community-detection algorithms for the automatic generation of ACLs in OSNs. Particularly, it analyses the performance of the aforementioned approach under different homophily conditions through a simulation model. Furthermore, since private information may reach the scope of untrusted recipients through the re-sharing affordances of OSNs, information diffusion processes are also modelled and taken explicitly into account. Altogether, the removal of gatekeeper nodes is further explored as a strategy to counteract unwanted data dissemination.

cs.SI↗

arXiv2018

Topic Modelling of Empirical Text Corpora: Validity, Reliability, and Reproducibility in Comparison to Semantic Maps

Using the 6,638 case descriptions of societal impact submitted for evaluation in the Research Excellence Framework (REF 2014), we replicate the topic model (Latent Dirichlet Allocation or LDA) made in this context and compare the results with factor-analytic results using a traditional word-document matrix (Principal Component Analysis or PCA). Removing a small fraction of documents from the sample, for example, has on average a much larger impact on LDA than on PCA-based models to the extent that the largest distortion in the case of PCA has less effect than the smallest distortion of LDA-based models. In terms of semantic coherence, however, LDA models outperform PCA-based models. The topic models inform us about the statistical properties of the document sets under study, but the results are statistical and should not be used for a semantic interpretation - for example, in grant selections and micro-decision making, or scholarly work-without follow-up using domain-specific semantic maps.

cs.CL↗