Both Text and Images Leaked! A Systematic Analysis of Multimodal LLM Data Contamination Paper ā¢ 2411.03823 ā¢ Published 4 days ago ā¢ 41
Decoding Dark Matter: Specialized Sparse Autoencoders for Interpreting Rare Concepts in Foundation Models Paper ā¢ 2411.00743 ā¢ Published 8 days ago ā¢ 6
Hunyuan-Large: An Open-Source MoE Model with 52 Billion Activated Parameters by Tencent Paper ā¢ 2411.02265 ā¢ Published 5 days ago ā¢ 22
LIBMoE: A Library for comprehensive benchmarking Mixture of Experts in Large Language Models Paper ā¢ 2411.00918 ā¢ Published 9 days ago ā¢ 8
How Far is Video Generation from World Model: A Physical Law Perspective Paper ā¢ 2411.02385 ā¢ Published 5 days ago ā¢ 27
DynaMath: A Dynamic Visual Benchmark for Evaluating Mathematical Reasoning Robustness of Vision Language Models Paper ā¢ 2411.00836 ā¢ Published 11 days ago ā¢ 14
AAAR-1.0: Assessing AI's Potential to Assist Research Paper ā¢ 2410.22394 ā¢ Published 11 days ago ā¢ 13
BitStack: Fine-Grained Size Control for Compressed Large Language Models in Variable Memory Environments Paper ā¢ 2410.23918 ā¢ Published 10 days ago ā¢ 17
GlotCC: An Open Broad-Coverage CommonCrawl Corpus and Pipeline for Minority Languages Paper ā¢ 2410.23825 ā¢ Published 10 days ago ā¢ 3
Language Models can Self-Lengthen to Generate Long Texts Paper ā¢ 2410.23933 ā¢ Published 10 days ago ā¢ 15
A Pointer Network-based Approach for Joint Extraction and Detection of Multi-Label Multi-Class Intents Paper ā¢ 2410.22476 ā¢ Published 11 days ago ā¢ 24
VideoWebArena: Evaluating Long Context Multimodal Agents with Video Understanding Web Tasks Paper ā¢ 2410.19100 ā¢ Published 16 days ago ā¢ 6
Constraint Back-translation Improves Complex Instruction Following of Large Language Models Paper ā¢ 2410.24175 ā¢ Published 9 days ago ā¢ 15
On Memorization of Large Language Models in Logical Reasoning Paper ā¢ 2410.23123 ā¢ Published 11 days ago ā¢ 15
ReferEverything: Towards Segmenting Everything We Can Speak of in Videos Paper ā¢ 2410.23287 ā¢ Published 10 days ago ā¢ 17
A Large Recurrent Action Model: xLSTM enables Fast Inference for Robotics Tasks Paper ā¢ 2410.22391 ā¢ Published 11 days ago ā¢ 21
AQLM+PV Collection Official AQLM quantizations for "PV-Tuning: Beyond Straight-Through Estimation for Extreme LLM Compression": https://arxiv.org/abs/2405.14852 ā¢ 25 items ā¢ Updated 1 day ago ā¢ 18