← Back to papers

2024 ICML ICML 2024

Eureka-Moments in Transformers: Multi-Step Tasks Reveal Softmax Induced Optimization Problems

Authors

David T Hoffmann , Simon Schrodi , Jelena Bratulić , Nadine Behrmann , Volker Fischer , Thomas Brox

Related papers

Learning Latent Dynamic Robust Representations for World Models 2024

Beyond Individual Input for Deep Anomaly Detection on Tabular Data 2024

Risk Estimation in a Markov Cost Process: Lower and Upper Bounds 2024

Collapse-Aware Triplet Decoupling for Adversarially Robust Image Retrieval 2024

Ranking-based Client Imitation Selection for Efficient Federated Learning 2024