AI研究論文每日精選

CORAL: Benchmarking Multi-turn Conversational Retrieval-Augmentation Generation

TokenFormer: Rethinking Transformer Scaling with Tokenized Model Parameters

A Large Recurrent Action Model: xLSTM enables Fast Inference for Robotics Tasks

ReferEverything: Towards Segmenting Everything We Can Speak of in Videos

On Memorization of Large Language Models in Logical Reasoning

Decoding Reading Goals from Eye Movements

Stealing User Prompts from Mixture of Experts

Toxicity of the Commons: Curating Open-Source Pre-Training Data

SlowFast-VGen: Slow-Fast Learning for Action-Driven Long Video Generation

Can Models Help Us Create Better Models? Evaluating LLMs as Data Scientists

AutoMIR: Effective Zero-Shot Medical Information Retrieval without Relevance Labels