Beyond fine-tuning: LoRA modules boost near-OOD detection and LLM security
Related publications (32)
Graph Chatbot
Chat with Graph Search
Ask any question about EPFL courses, lectures, exercises, research, news, etc. or try the example questions below.
DISCLAIMER: The Graph Chatbot is not programmed to provide explicit or categorical answers to your questions. Rather, it transforms your questions into API requests that are distributed across the various IT services officially administered by EPFL. Its purpose is solely to collect and recommend relevant references to content that you can explore to help you answer your questions.
In this paper, we investigate the construction of compromise estimators of location and scale, by averaging over several models selected among a specified large set of possible models. The weight given to each distribution is based on the profile likelihoo ...
The idea of stochastic sediment transport models emerged in the 1930s, notably with the doctoral work of Hans A. Einstein (1936). Einstein's seminal work gave impulse to several stochastic models, which usually led to thin-tailed or bounded distributions f ...
The thesis is a contribution to extreme-value statistics, more precisely to the estimation of clustering characteristics of extreme values. One summary measure of the tendency to form groups is the inverse average cluster size. In extreme-value context, th ...
To aid assessments of climate change impacts on water related activities in the case study regions (CSRs) of the EC funded project SWURVE, estimates of uncertainty in climate model data need to be developed. In this paper, two methods to estimate uncertain ...
The evaluation of avalanche release depth distributions represents a major challenge for hazard management. This paper presents a rigorous formalism in which these distributions are expressed through a coupling of mechanical and meteorological factors. Con ...
Nanoparticles are finding many research and industrial applications, yet their characterization remains a challenge. Their cores are often polydisperse and coated by a stabilizing shell that varies in size and composition. No single technique can character ...
In this paper, we discuss some of the issues that arise with the computation of the implied value of travel-time savings in the case of discrete choice models allowing for random taste heterogeneity. We specifically look at the case of models producing a n ...
We present Scanning Mobility CCN Analysis (SMCA) as a novel method for obtaining rapid measurements of size-resolved cloud condensation nuclei (CCN) distributions and activation kinetics. SMCA involves sampling the monodisperse outlet stream of a Different ...
This paper presents a new algorithm for classifying distributions. The algorithm combines the principle of margin maximization and a kernel trick, applied to distributions. Thus, it combines the discriminative power of support vector machines and the well- ...
We present a general method for maintaining estimates of the distribution of parameters in arbitrary models. This is then applied to the estimation of probability distribution over actions in value-based reinforcement learning. While this approach is simil ...