Neural Network Optimization

Taming LLMs with Gradient Grouping ACL 2025 Targeted Low-rank Refinement: Enhancing Sparse Language Models with Precision ICML 2025 Task Descriptors Help Transformers Learn Linear Models In-Context ICLR 2025 Task-driven Layerwise Additive Activation Intervention NAACL 2025 Task-Specific Gradient Adaptation for Few-Shot One-Class Classification CVPR 2025 Teaching Models to Improve on Tape AAAI 2025 TETRIS: Optimal Draft Token Selection for Batch Speculative Decoding ACL 2025 Text2Sql: Pure Fine-Tuning and Pure Knowledge Distillation NAACL 2025 Text Detoxification: Data Efficiency, Semantic Preservation and Model Generalization EMNLP 2025 \textttconfopt: A Library for Implementation and Evaluation of Gradient-based One-Shot NAS Methods AutoML 2025 The Directionality of Optimization Trajectories in Neural Networks ICLR 2025 The Effect of SGD Batch Size on Autoencoder Learning: Sparsity, Sharpness, and Feature Learning JMLR 2025 The Gradient of Algebraic Model Counting AAAI 2025 The Hidden Attention of Mamba Models ACL 2025 The Optimization Landscape of SGD Across the Feature Learning Strength ICLR 2025 Theoretical Insights into Fine-Tuning Attention Mechanism: Generalization and Optimization IJCAI 2025 The Power of Random Features and the Limits of Distribution-Free Gradient Descent ICML 2025 The Role of Outgoing Connection Heterogeneity in Feedforward Layers of Large Language Models EMNLP 2025 The Surprising Effectiveness of Infinite-Width NTKs for Characterizing and Improving Model Training AAAI 2025 Think Clearly: Improving Reasoning via Redundant Token Pruning EMNLP 2025 Thinking Before You Speak: A Proactive Test-time Scaling Approach EMNLP 2025 ThinkSwitcher: When to Think Hard, When to Think Fast EMNLP 2025 ThinkTuning: Instilling Cognitive Reflections without Distillation EMNLP 2025 ThoughtProbe: Classifier-Guided LLM Thought Space Exploration via Probing Representations EMNLP 2025 Three Mechanisms of Feature Learning in a Linear Network ICLR 2025