NeurIPS 2026 reviews are up: reviews, scores, rebuttals and meta-reviews for 7,877 papers. See the highest rated
acceptodds: the papers of NeurIPS 2026
Abstract
The safe for work edition of acceptodds: the submissions, without the markets. Search them, browse them on a map laid out by what they are about, follow the ones you care about, keep reading lists with your group, and get recommendations from what you read. There are no odds, no bets and no reputation here. acceptodds is not affiliated with any organisation or institution. acceptodds is an open source project, open to contributions, at github.com/PotentialHedge/acceptodds-sfw.1 Follow for updates on X.
[PDF]Concept Modulation Models: A Unified Framework for Identifiability and ExtrapolationSoheun Yi, Yizhou Lu, Chandler Squires et al.4.33
[abs]Safe Actions Can Form Unsafe Traces: Benchmarking and Shielding Compositional Emergent Risk in AI AgentsZhongze Wu, Xiu Su, Hongyan Xu et al.4.33
[abs]ProjKAN: Model Compression via KAN Projections to Bridge the Hypothesis and Capacity GapsFerhat Arslan, Weihong Guo, Shuo Li4.33
[abs]Diffusion-Time Concept Manifolds: Sparse Autoencoder Groups for Interpreting Denoising Language ModelsPawan Kumar4.33
[abs]Horizontal Diffusion Models: Score-based Generative Modeling on Frame-Connection GeometrySungwoo Park4.33
[abs]Adversarially Attacking Symbolic Vocabulary Vulnerabilities In LLM PlannersStephen Obadinma, Xiaodan Zhu4.33
[PDF]Active Learning From Positive and Unlabeled ExamplesFarnam Mansouri, Sandra Zilles, Shai Ben-David4.33
[abs]Calibrated Target Noise Recovers Curvature from the GradientsArash Jamshidi, Katsiaryna Haitsiukevich, Aristides Gionis et al.4.33
[PDF]How Accurately Can a Gaussian Approximate Stochastic Approximation Iterates?Shaan U Haque, Zedong Wang, Zixuan Zhang et al.4.33
[abs]An Equivariance Principle for Optimizer Design: Symmetry-Compatible Updates for Embeddings, LM Heads, and MoE RoutersTim Tsz-Kit Lau, Weijie Su4.33
[PDF]Mechanistic Circuit Identification for Controllable Data GenerationNakyung Lee, Sangwoo Hong, Jungwoo Lee4.33
[PDF]Learning Chance-Constrained MDPs with Bellman Distributional CertificatesChenbei Lu, Hongyu Yi4.33
[PDF]AEM: Adaptive Entropy Modulation for Multi-Turn Agentic Reinforcement LearningHaotian Zhao, Songlin Zhou, Yuxin Zhang et al.4.33
[PDF]fmxcoders: Factorized Masked Crosscoders for Cross-Layer Feature DiscoveryAndreas D Demou, Panagiotis Koromilas, James Oldfield et al.4.33
[PDF]DROGO: Default Representation Objective via Graph Optimization in Reinforcement LearningHon Tik (Rick) Tse, Marlos C. Machado4.33
[abs]Reasoning Pathologies in Large Language Models: A Diagnostic PerspectiveRohan Surana, Junda Wu, Sheldon Yu et al.4.33
[abs]Effective Knowledge Conflict Detection via Joint Agentic OptimizationYi Hong, Wenchao Bai, Jiaqi Jiang et al.4.33
[abs]Informed Posterior Sampling: More Efficient Online Learning with Few Offline Demonstrations in Average-Reward MDPsDengwang Tang, Rahul Jain, Botao Hao et al.4.33
[PDF]Posterior-First Neural PDE Simulation: Inferring Hidden Problem State from a Single FieldWenshuo Wang, Fan Zhang4.33
[PDF]GOLD PANNING: Strategic Context Shuffling for Needle-in-Haystack ReasoningAdam Byerly, Daniel Khashabi4.33
[abs]TReDS: Trajectory-Grounded Requirement-Capability Modeling for Training Distribution Shaping in Tool-Interactive TasksHan Liu4.33
[PDF]AtomWorld-Mem: Memory-Restored World States for Long-Horizon Atomistic EvolutionTian Luo, Ruge Zhang, Haozhi Han et al.4.33
[abs]PE-SHAP: Causally Interpretable Path-Wise Shapley ExplanationsMaksym Buleshnyi, Joshua Loftus, Sakina Hansen4.33
[abs]Scale-mixture Langevin sampling in the subspaces of recurrent cortical circuit dynamicsZimei Chen, Yi Ren, Wen-Hao Zhang4.33
[PDF]SOLAR: AI-Powered Speed-of-Light Performance AnalysisQijing Huang, Sana Damani, Zhifan Ye et al.4.33
[abs]GAUGECAST++: A Physics-Informed Latent Forecasting System for Localized Flood PredictionAdrian Kucia, Edward Rollason, Wai Lok Woo4.33
[abs]SLiDE: Structured Linear Dynamics for Forecasting with Exogenous InputsSebastian Pütz, Theodore Glavas, Benjamin Schäfer et al.4.33
[PDF]FiLoRA: Focus-and-Ignore LoRA for Controllable Feature RelianceHyunsuk Chung, Caren Han, Seungyeon Ji et al.4.33
[PDF]Optimal and Efficient Contextual Combinatorial Semi-bandits with General Function ApproximationHao Qin, Chicheng Zhang4.33
[abs]Leakage Thresholds for Sandwich Equilibria Under Partial InformationTingyi Lin, Ruoran Lai, Jiazhuo Li4.33
[PDF]Neural Networks With Dense Weights Are Not Universal ApproximatorsLevi Rauchwerger, Stefanie Jegelka, Ron Levie4.33
[PDF]When Dynamics Shift, Robust Task Inference Wins: Offline Imitation Learning with Behavior Foundation Models RevisitedRishabh Agrawal, Rahul Jain, Ashutosh Nayyar4.33
[abs]What Arranges Features in Activation Space? Non-Classical Predictive Geometry in Next-Token PredictorsAdam Shai, Thomas J Elliott, Paul Riechers4.33
[abs]GraphUAT: Uncertainty Attribution in Graph Neural NetworksBurouj Armgaan, Karan R Singh, Abhirup Mondal et al.4.33
[abs]Finite-Sample Convergence in Networked Average Reward MARL: Decentralization Pitfalls and Entropy RemediesYizhou Zhang, Yashaswini Murthy, Laixi Shi et al.4.33
[abs]Beyond Expected Values: Risk-Sensitive Planning with Distributional Monte-Carlo Tree SearchQuynh Trinh, Trung Nguyen, Nam Nguyen et al.4.33
[abs]UGM: Unified Multi-scale Genomic Event Modeling with Site-level Joint PredictionJiayang Wu, Chenchen Qin, Xu YANG et al.4.33
[abs]SeqDiCO: Sequence-oriented Diffusion for Scale-Generalizable Neural Combinatorial OptimizationYu Wang, Qiaolin Lu, Yang Wu et al.4.33
[PDF]QMaxCal: Path-Space Regularization for Open Quantum Control via Girsanov's TheoremMerijn Moody, Zier Mensch, Miranda Cheng et al.4.33
[PDF]MAGE: Multi-Agent Self-Evolution with Co-Evolutionary Knowledge GraphsRuiyi Yang, Zechen Li, Hao Xue et al.4.33
[abs]Locally-Additive Regret for Delayed Non-stationary Bandit Convex OptimizationWentao Zhang, Yifan Zhu, Yutong Zhang4.33
[abs]DUST: Directional Uncertainty-aware and Scale-invariant Transfer LearningAbhisek Chakraborty4.33
1 If this project is interesting to you, reach out at [email protected].