3 papers
cs.LG2025
Towards Safeguarding LLM Fine-tuning APIs against Cipher Attacks
Jack Youstra, Mohammed Mahfoud, Yang Yan +3
Large language model fine-tuning APIs enable widespread model customization, yet pose significant safety risks. Recent work shows that adversaries can exploit access to these APIs…
cs.LG2025
Learning Decision Trees as Amortized Structure Inference
Mohammed Mahfoud, Ghait Boukachab, Michał Koziarski +4
Building predictive models for tabular data presents fundamental challenges, notably in scaling consistently, i.e., more resources translating to better performance, and generalizi…
cs.LG2025
Forecasting Rare Language Model Behaviors
Erik Jones, Meg Tong, Jesse Mu +7
Standard language model evaluations can fail to capture risks that emerge only at deployment scale. For example, a model may produce safe responses during a small-scale beta test,…