From the 1 of 5 linked papers with an AI index.
5 papers
Test-Time Augmentation for LLMs: When Input Diversity Beats Output Diversity at Matched Compute
Nikita Kozodoi, Zainab Afolabi, Jack Butler
Test-time scaling improves LLM accuracy but multiplies inference cost, making the accuracy gained per unit of compute the metric that matters in deployment. Self-consistency is one…
Are we Merging the Right Models? Impact of Expert Training Duration on Model Merging for LLMs
Nikita Kozodoi, Zainab Afolabi, Jack Butler
The paper investigates how the training duration of domain-specific expert models influences the performance of merged large language models, finding that optimal merging strategie…
SWE-InfraBench: Evaluating Language Models on Cloud Infrastructure Code
Natalia Tarasova, Enrique Balp-Straffon, Aleksei Iancheruk +10
Building infrastructure-as-code (IaC) in cloud computing is a critical task, underpinning the reliability, scalability, and security of modern software systems. Despite the remarka…
Finding the Sweet Spot: Trading Quality, Cost, and Speed During Inference-Time LLM Reflection
Jack Butler, Nikita Kozodoi, Zainab Afolabi +2
As Large Language Models (LLMs) continue to evolve, practitioners face increasing options for enhancing inference-time performance without model retraining, including budget tuning…
Fighting Sampling Bias: A Framework for Training and Evaluating Credit Scoring Models
Nikita Kozodoi, Stefan Lessmann, Morteza Alamgir +2
Scoring models support decision-making in financial institutions. Their estimation and evaluation are based on the data of previously accepted applicants with known repayment behav…