MÖVE: New Benchmark for Evaluating Large Language Models in German Public Sector
Researchers have created MÖVE, a comprehensive benchmark designed to evaluate large language models specifically for use in German public administration. The benchmark assesses 39 models across performance metrics (summarization, question answering, topic extraction) and governance criteria (hallucinations, energy consumption, constitutional alignment). This addresses a gap in existing benchmarks, which are predominantly English and US-centric, and provides guidance for model selection in public sector contexts.
MÖVE (Modelle für die Öffentliche Verwaltung Evaluieren) is a holistic evaluation framework for large language models tailored to the German public sector. The benchmark evaluates 39 models using ten German-language datasets, including newly constructed gold- and silver-standard datasets reflecting public administration domains. Beyond traditional performance metrics like summarization and question answering, MÖVE uniquely incorporates governance criteria including hallucination assessment, energy consumption measurement, provider transparency, and alignment with German constitutional values and knowledge of German political party positions. The research employs multiple evaluation methodologies combining classical NLP metrics, embedding-based methods, and LLM-as-a-judge approaches. Key findings indicate that no single model excels across all criteria, with top performers varying by task, and that model size alone is a poor predictor of quality. The benchmark itself underwent rigorous validation examining statistical precision, LLM judge reliability, dataset impact on rankings, prompt sensitivity, and energy consumption estimate validity.
What different sources said
- arXiv cs.CLCenter
M\"OVE: A Holistic LLM Benchmark for the German Public Sector
Related
Topology-Aware Thermodynamics Improves DNA Probe Specificity Design
Researchers developed a new framework for designing DNA probes that accounts for the spatial organization of matched sequences, not just overall thermodynamic stability. Traditional methods rely on scalar measures like melting temperature and free energy, which miss how mismatches are distributed along the probe. The approach could improve diagnostic accuracy in applications like HPV detection and gene expression profiling.
Study Identifies Optimal Thermal Dose for Combining Focused Ultrasound with Immunotherapy in Tumors
Researchers used multimodal PET imaging to identify an optimal thermal dose range for focused ultrasound ablation that destroys tumor tissue while preserving conditions for immunotherapy delivery. The study found that excessive heating collapses blood vessels needed for antibody access, while insufficient heating fails to adequately reduce tumor burden. The findings could guide clinical design of combination treatments pairing thermal ablation with immunotherapies.
Plant MSH1 Protein Functions as Mismatch-Directed Nuclease for Organelle Genome Maintenance
Researchers have identified the precise mechanism by which the AtMSH1 protein in Arabidopsis plants recognizes and cleaves DNA mismatches and lesions, preventing mutations in organellar genomes. The protein combines a DNA mismatch recognition module with a nuclease domain that makes staggered cuts at specific positions relative to DNA damage. This discovery explains how plants maintain unusually low mutation rates in their mitochondrial and chloroplast DNA compared to other eukaryotes.