Modality Gap-Driven Subspace Alignment Training Paradigm For Multimodal Large Language Models Paper • 2602.07026 • Published 8 days ago • 108 • 4
Learning a Generative Meta-Model of LLM Activations Paper • 2602.06964 • Published 4 days ago • 2 • 3
view article Article From Golden Gate Bridge to Broken JSON: Why Anthropic's SAE Steering Fails for Structured Output 3 days ago • 14
view article Article Training Design for Text-to-Image Models: Lessons from Ablations 7 days ago • 55
Understanding self-supervised Learning Dynamics without Contrastive Pairs Paper • 2102.06810 • Published Feb 12, 2021 • 1
view article Article Introducing Daggr: Chain apps programmatically, inspect visually +3 13 days ago • 94
Shaping capabilities with token-level data filtering Paper • 2601.21571 • Published 12 days ago • 25 • 3