Sep 2026· Zenodo (CERN European Organization for Nuclear Research)
Abstract
This deposit contains the prospectively frozen ARS Cross-Domain Evaluation Protocol v1.8 and its reproducibility artifacts. The protocol evaluates whether the Adaptive Research Scoping (ARS) four-dimensional corpus qualification architecture generalizes across deliberately different evidence domains while allowing only the operationalization of A (Alignment), C (Coverage), K (Contamination), and R (Contextual Representation) to vary by domain. Target domains: 1. Biomedical evidence 2. Education 3. Psychology 4. Computer science / artificial intelligence 5. Social science The package contains: - ARS v1.8 protocol and preregistration-ready statistical analysis plan - reproducible analysis scaffold - frozen configuration template - README and execution instructions - SHA-256 freeze manifest - cryptographic freeze-proof JSON The SHA-256 manifest is intended to provide content-integrity evidence. The local cryptographic freeze should not be interpreted as an independent trusted timestamp. This Zenodo deposit provides an independent archival record and publication timestamp. The protocol explicitly prohibits silently substituting missing domains. A five-domain generalization claim is permitted only if all five target domains meet the predefined eligibility criteria.
GAOKAO-Bench is introduced, an intuitive benchmark that employs questions from the Chinese GAOKAO examination as test samples, including both subjective and objective questions that contribute a robust evaluation benchmark for future large language models and offers valuable insights into the advantages and limitations of such models.
Xiaotian Zhang, Chun-yan Li, Yi Zong et al.· arXiv.org· 216 citations· ⚡17
Empirically, PRISM reduces the end-to-end time for data selection and model tuning to just 30% of conventional pipelines, and achieves this efficiency while simultaneously enhancing performance, surpassing models fine-tuned on the full dataset across eight multimodal and three language understanding benchmarks.
Jinhe Bi, Yifan Wang, Danqi Yan et al.· arXiv.org· 73 citations· ⚡4
The method, ECCOLA, is presented, which aims at making the high-level AI ethics principles more practical, making it possible for developers to more easily implement them in practice.
Ville Vakkuri, Kai-Kristian Kemell, P. Abrahamsson· EUROMICRO Conference on Soft...· 64 citations· ⚡6
This paper designs Markov decision processes (MDPs) for different combinatorial problems and proposes to train conditional GFlowNets to sample from the solution space and demonstrates that GFlowNet policies can efficiently find high-quality solutions.
Dinghuai Zhang, H. Dai, Esmeralda S. Whitammer et al.· Advances in Neural Informati...· 59 citations· ⚡8
An empirical study on the current state of practice in artificial intelligence ethics is conducted by means of a multiple case study of five case companies, which indicates a gap between research and practice in the area.
Ville Vakkuri, Kai-Kristian Kemell, Joni Kultanen et al.· arXiv.org· 56 citations· ⚡6