training dynamics

Identifying Noise in Human-Created Datasets using Training Dynamics from Generative Models EMNLP 2025 Linear Steerability in Language Models: When It Emerges and How It Evolves EMNLP 2025 Mirror, Mirror of the Flow: How Does Regularization Shape Implicit Bias? ICML 2025 On the Optimization and Generalization of Two-layer Transformers with Sign Gradient Descent ICLR 2025 Position: An Empirically Grounded Identifiability Theory Will Accelerate Self Supervised Learning Research ICML 2025 Random Pruning Over-parameterized Neural Networks Can Improve Generalization: A Training Dynamics Analysis JMLR 2025 Rethinking Benign Overfitting in Two-Layer Neural Networks ICML 2025 Simulating Training Dynamics to Reconstruct Training Data from Deep Neural Networks ICLR 2025 SOI Matters: Analyzing Multi-Setting Training Dynamics in Pretrained Language Models via Subsets of Interest EMNLP 2025 The Surprising Effectiveness of Infinite-Width NTKs for Characterizing and Improving Model Training AAAI 2025 Token-Level Metrics for Detecting Incorrect Gold Annotations in Named Entity Recognition EMNLP 2025 Training Dynamics Underlying Language Model Scaling Laws: Loss Deceleration and Zero-Sum Learning ACL 2025 Understanding Silent Data Corruption in LLM Training ACL 2025 When Can We Approximate Wide Contrastive Models with Neural Tangent Kernels and Principal Component Analysis? AAAI 2025 Your Model is Overconfident, and Other Lies We Tell Ourselves ACL 2025 Abrupt Learning in Transformers: A Case Study on Matrix Completion NeurIPS 2024 Analyzing and Improving the Training Dynamics of Diffusion Models CVPR 2024 An exactly solvable model for emergence and scaling laws in the multitask sparse parity problem NeurIPS 2024 Causal Estimation of Memorisation Profiles ACL 2024 CLUES: Collaborative Private-domain High-quality Data Selection for LLMs via Training Dynamics NeurIPS 2024 Critical Learning Periods: Leveraging Early Training Dynamics for Efficient Data Pruning ACL 2024 Deep Fusion: Efficient Network Training via Pre-trained Initializations ICML 2024 Enhancing Neural Training via a Correlated Dynamics Model ICLR 2024 Evaluating the design space of diffusion-based generative models NeurIPS 2024 Grokking in Linear Estimators -- A Solvable Model that Groks without Understanding ICLR 2024