4 papers
Residual Connections and the Causal Shift: Uncovering a Structural Misalignment in Transformers
Jonathan Lys, Vincent Gripon, Bastien Pasdeloup +4
Large Language Models (LLMs) are trained with next-token prediction, implemented in autoregressive Transformers via causal masking for parallelism. This creates a subtle misalignme…
Inner Loop Inference for Pretrained Transformers: Unlocking Latent Capabilities Without Training
Jonathan Lys, Vincent Gripon, Bastien Pasdeloup +4
Deep Learning architectures, and in particular Transformers, are conventionally viewed as a composition of layers. These layers are actually often obtained as the sum of two contri…
GaLLoP: Gradient-based Sparse Learning on Low-Magnitude Parameters
Anand Choudhary, Yasser Sulaıman, Lukas Mauch +3
Sparse fine-tuning techniques adapt LLMs to downstream tasks by only tuning a sparse subset of model parameters. However, the effectiveness of sparse adaptation depends on optimall…
A Novel Benchmark for Few-Shot Semantic Segmentation in the Era of Foundation Models
Reda Bensaid, Vincent Gripon, François Leduc-Primeau +3
Few-shot semantic segmentation (FSS) is a crucial challenge in computer vision, driving extensive research into a diverse range of methods, from advanced meta-learning techniques t…