Papers published in 2025

• xLSTM 7B: A Recurrent LLM for Fast and Efficient Inference (presented at ICML 2025)

• Tiled Flash Linear Attention: More Efficient Linear RNN and xLSTM Kernels (Workshop on Foundation Models in the Wild – ICLR 2025)

• TiRex: Zero-Shot Forecasting Across Long and Short Horizons with Enhanced In-Context Learning (presented at NeurIPS 2025)

• TiRex Classification (presented at NeurIPS BERT2S WS 2025)

• xLSTM Scaling Laws: Competitive Performance with Linear Time-Complexity

Papers published in 2024

• xLSTM: Extended Long Short-Term Memory (presented at NeurIPS 2024)

• Vision-LSTM: xLSTM as Generic Vision Backbone (presented at ICLR 2025)

• A Large Recurrent Action Model: xLSTM enables Fast Inference for Robotics Tasks (presented at ICML 2025)

• Bio-xLSTM: Generative modeling, representation and in-context learning of biological and chemical sequences (presented at ICLR 2025)

• FlashRNN: I/O-Aware Optimization of Traditional RNNs on modern hardware (presented at ICLR 2025)