ICLR 2026 に論文が採択されました。
- Kazuki Yano, Shun Kiyono, Sosuke Kobayashi, Sho Takase, Jun Suzuki.
“Pre-training LLM without Learning Rate Decay Enhances Supervised Fine-Tuning”
- Taishi Nakamura, Satoki Ishikawa, Masaki Kawamura, Takumi Okamoto, Daisuke Nohara, Jun Suzuki, Rio Yokota.
“Optimal Sparsity of Mixture-of-Experts Language Models for Reasoning Tasks”