5 papers
Beyond Accuracy: Decomposing the Reasoning Efficiency of LLMs
Daniel Kaiser, Arnoldo Frigessi, Ali Ramezani-Kebrya +1
As reasoning LLMs increasingly trade tokens for accuracy through deliberation, search, and self-correction, a single accuracy score can no longer tell whether those tokens buy usef…
Stepwise Variational Inference with Vine Copulas
Elisabeth Griesbauer, Leiv Rønneberg, Arnoldo Frigessi +2
We propose stepwise variational inference (VI) with vine copulas: a universal VI procedure that combines vine copulas with a novel stepwise estimation procedure of the variational…
CogniLoad: A Synthetic Natural Language Reasoning Benchmark With Tunable Length, Intrinsic Difficulty, and Distractor Density
Daniel Kaiser, Arnoldo Frigessi, Ali Ramezani-Kebrya +1
Current benchmarks for long-context reasoning in Large Language Models (LLMs) often blur critical factors like intrinsic task complexity, distractor interference, and task length.…
TVineSynth: A Truncated C-Vine Copula Generator of Synthetic Tabular Data to Balance Privacy and Utility
Elisabeth Griesbauer, Claudia Czado, Arnoldo Frigessi +1
We propose TVineSynth, a vine copula based synthetic tabular data generator, which is designed to balance privacy and utility, using the vine tree structure and its truncation to d…
BoRA: Bayesian Hierarchical Low-Rank Adaption for Multi-Task Large Language Models
Simen Eide, Arnoldo Frigessi
This paper introduces Bayesian Hierarchical Low-Rank Adaption (BoRA), a novel method for finetuning multi-task Large Language Models (LLMs). Current finetuning approaches, such as…