{"total":34,"rows":[{"id":"2512.24880","title":"mHC: Manifold-Constrained Hyper-Connections","created":"2025-12-31","cited_by_count":88,"area":null},{"id":"2603.15031","title":"Attention Residuals","created":"2026-03-16","cited_by_count":65,"area":null},{"id":"2407.09298","title":"Transformer Layers as Painters","created":"2024-07-12","cited_by_count":54,"area":null},{"id":"2502.12170","title":"MUDDFormer: Breaking Residual Bottlenecks in Transformers via Multiway Dynamic Dense Connections","created":"2025-02-13","cited_by_count":31,"area":null},{"id":"2410.24159","title":"GPT or BERT: why not both?","created":"2024-10-31","cited_by_count":27,"area":null},{"id":"2601.21582","title":"Depth-Recurrent Attention Mixtures: Giving Latent Reasoning the Attention it Deserves","created":"2026-01-29","cited_by_count":12,"area":null},{"id":"2604.21254","title":"Hyperloop Transformers","created":"2026-04-23","cited_by_count":12,"area":null},{"id":"2601.00417","title":"Deep Delta Learning","created":"2026-01-01","cited_by_count":9,"area":null},{"id":"2502.06785","title":"DeepCrossAttention: Supercharging Transformer Residual Connections","created":"2025-02-10","cited_by_count":8,"area":null},{"id":"2509.18993","title":"CR-Net: Scaling Parameter-Efficient Training with Cross-Layer Low-Rank Structure","created":"2025-09-23","cited_by_count":7,"area":null},{"id":"2605.18855","title":"Delta Attention Residuals","created":"2026-05-13","cited_by_count":7,"area":null},{"id":"2603.15619","title":"Mixture-of-Depths Attention","created":"2026-03-16","cited_by_count":6,"area":null},{"id":"2502.09245","title":"You Do Not Fully Utilize Transformer's Representation Capacity","created":"2025-02-13","cited_by_count":4,"area":null},{"id":"2506.09714","title":"Auto-Compressing Networks","created":"2025-06-11","cited_by_count":4,"area":null},{"id":"2604.02309","title":"go-$m$HC: Direct Parameterization of Manifold-Constrained Hyper-Connections via Generalized Orthostochastic Matrices","created":"2026-04-02","cited_by_count":4,"area":null},{"id":"2603.16039","title":"Residual Stream Duality in Modern Transformer Architectures","created":"2026-03-17","cited_by_count":3,"area":null},{"id":"2607.27230","title":"Multi-Head Attention Residuals","created":"2026-07-22","cited_by_count":2,"area":null},{"id":"2601.22095","title":"GeoNorm: Unify Pre-Norm and Post-Norm with Geodesic Optimization","created":"2026-01-29","cited_by_count":1,"area":null},{"id":"2603.22665","title":"Improving LLM Final Representations with Inter-Layer Geometry","created":"2026-03-24","cited_by_count":1,"area":null},{"id":"2605.03953","title":"Transformers with Selective Access to Early Representations","created":"2026-05-05","cited_by_count":1,"area":null},{"id":"2605.20798","title":"Most Transformer Modifications Still Do Not Transfer at 1-3B: A 2020-2026 Update to Narang et al. (2021) with Downstream Evaluation and a Noise Floor","created":"2026-05-20","cited_by_count":1,"area":null},{"id":"2606.02780","title":"Do Value Vectors in Deep Layers Need Context from the Residual Stream?","created":"2026-06-01","cited_by_count":1,"area":null},{"id":"2607.09694","title":"Low-Rank Attention Residuals","created":"2026-06-19","cited_by_count":1,"area":null},{"id":"2607.18730","title":"Dual Attention Residuals","created":"2026-07-21","cited_by_count":1,"area":null},{"id":"2609.32534","title":"DepthBench: Measuring How Residual Connections Enable More Computational Depth","created":"2026-09-26","cited_by_count":1,"area":null},{"id":"2606.06564","title":"HAARES Half-Split Residual Basis Routing for Deep Transformers","created":"2026-06-04","cited_by_count":0,"area":null},{"id":"2607.26192","title":"Dynamic Parameterization Is Not Dynamic Inference","created":"2026-07-28","cited_by_count":0,"area":null},{"id":"2608.01075","title":"Role-Decoupled Attention Residuals: Separating Matching and Content Retrieval Across Depth","created":"2026-08-02","cited_by_count":0,"area":null},{"id":"2608.18486","title":"WhiteMatter: All-to-All Cross-Layer Connections via KV Source Mixing","created":"2026-08-19","cited_by_count":0,"area":null},{"id":"2609.10715","title":"NCP-ArchPreview Technical Report: Moving towards Latent Space Language Models through Next Concept Prediction","created":"2026-09-09","cited_by_count":0,"area":null},{"id":"2609.39109","title":"T-Router: Learning Thalamic Routing for Reasoning with Parameter-Efficient Reinforcement Learning","created":"2026-09-30","cited_by_count":0,"area":null},{"id":"2610.13630","title":"Multi-Stream Attention Residuals","created":"2026-10-09","cited_by_count":0,"area":null},{"id":"2610.18809","title":"Depth Memory","created":"2026-10-13","cited_by_count":0,"area":null},{"id":"2610.19625","title":"DAGFormer: Global Topology Prediction for Transformer Cross-Layer Connections","created":"2026-10-14","cited_by_count":0,"area":null}],"next_cursor":null}