arxiv:2609.37169
Changxin Tian
ChangxinTian
AI & ML interests
None yet
Recent Activity
authored a paper 2 days ago
Hyperparameter Scaling Laws Across MoE Sparsity authored a paper 2 days ago
Trajectory Soup: Pushing the Compute-Scaling Frontier of LLM Mid-training via Diverse Trajectories authored a paper 2 days ago
Effective Learning Rate Governs Loss Dynamics in Language Model PretrainingOrganizations
models 0
None public yet
datasets 0
None public yet