On Counterfactual Interventions in Vector Autoregressive Modelsby Kurt Butler, Marija Iloska, Petar M. DjuricFirst submitted…
Efficient World Models with Context-Aware Tokenizationby Vincent Micheli, Eloi Alonso, François FleuretFirst submitted to arxiv…
Sparser is Faster and Less is More: Efficient Sparse Attention for Long-Range Transformersby Chao Lou,…
Understanding and Mitigating Tokenization Bias in Language Modelsby Buu Phan, Marton Havasi, Matthew Muckley, Karen…
Why LLMs Are Bad at Synthetic Table Generation (and what to do about it)by Shengzhe…
HoTPP Benchmark: Are We Good at the Long Horizon Events Forecasting?by Ivan Karpukhin, Foma Shipilov,…
LTSM-Bundle: A Toolbox and Benchmark on Large Language Models for Time Series Forecastingby Yu-Neng Chuang,…
Probing the Emergence of Cross-lingual Alignment during LLM Trainingby Hetong Wang, Pasquale Minervini, Edoardo M.…
Transcendence: Generative Models Can Outperform The Experts That Train Themby Edwin Zhang, Vincent Zhu, Naomi…
Lift Your Molecules: Molecular Graph Generation in Latent Euclidean Spaceby Mohamed Amine Ketata, Nicholas Gao,…