3 papers
cs.AI2026
Multilingual GSM-Symbolic: What determines capability transfer across languages?
Kenneth Enevoldsen, Riley Herchert, Sofie Mosegaard +22
We understand little about how capabilities acquired in one language carry over to another, or what governs this transfer: evaluations rely on incomparable, saturation-prone datase…
cs.AI2026
Conformity Mitigations in Large Language Models Lie on a Single Resistance-Receptivity Frontier
Zafar Hussain, Kristoffer Nielbo
Recent advances in language models have enabled collaborative settings in which multiple models leverage one another's capabilities, iteratively improving, transforming, and extend…
cs.CL2026
The Coverage Illusion: From Pre-retrieval Routing Failure to Post-retrieval Cascades in a Production RAG System
Zafar Hussain, Kristoffer Nielbo
In modern RAG pipelines, query augmentation methods such as HyDE and query expansion are applied to every query, resulting in substantial LLM inference costs and increased end-to-e…