MIG: Automatic Data Selection for Instruction Tuning by Maximizing Information Gain in Semantic Space Paper • 2504.13835 • Published 5 days ago • 35
PEBench: A Fictitious Dataset to Benchmark Machine Unlearning for Multimodal Large Language Models Paper • 2503.12545 • Published Mar 16 • 5
Creation-MMBench: Assessing Context-Aware Creative Intelligence in MLLM Paper • 2503.14478 • Published Mar 18 • 47
MM-Eureka: Exploring Visual Aha Moment with Rule-based Large-scale Reinforcement Learning Paper • 2503.07365 • Published Mar 10 • 60
MM-Eureka: Exploring Visual Aha Moment with Rule-based Large-scale Reinforcement Learning Paper • 2503.07365 • Published Mar 10 • 60 • 2