List of papers covered:
Transformers are SSMs: Generalized Models and Efficient Algorithms Through Structured State Space Duality
MMLU-Pro: A More Robust and Challenging Multi-Task Language Understanding Benchmark (special appearance by authors)
Improving the Training of Rectified Flows
Are Long-LLMs A Necessity For Long-Context Tasks?
Zipper: A Multi-Tower Decoder Architecture for Fusing Modalities
FineWeb: decanting the web for the finest text data at scale
Offline Regularised Reinforcement Learning for Large Language Models Alignment
Transformers Can Do Arithmetic with the Right Embeddings
Guiding a Diffusion Model with a Bad Version of Itself
Self-Improving Robust Preference Optimization
Situational Awareness
AI papers of the week - June 5th, 2024 - MMLU-Pro, Zipper, FineWeb, etc.