Selective Training for Large Vision Language Models via Visual Information Gain Paper • 2602.17186 • Published 7 days ago • 2