Song, B., Peng, J., Zhang, Y., Chen, G., Yang, F., & Guo, J. (2026). KTV: Keyframes and Key Tokens Selection for Efficient Training-Free Video LLMs. Proceedings of the AAAI Conference on Artificial Intelligence, 40(11), 9060–9068. https://doi.org/10.1609/aaai.v40i11.37862