NeurIPS 2026 reviews are up: reviews, scores, rebuttals and meta-reviews for 7,877 papers. See the highest rated
acceptodds: the papers of ICLR 2027
7,673 results for “keyword:"ai"” in ICLR 2027. 1,836 withdrawn or desk rejected listed apart. clearsort:relevanceratingnewestmost read
[PDF]The Life Cycle of a Massive Activation: Stochastic Birth, Weight-Decay-Driven Growth, and Competitive Consolidation▼4.3k2.00
[PDF]DORIAN: Zero-Shot Directionally Structured Noise Removal via Graph Neural Priors in Fourier Space▼4.3k2.00
[PDF]The Attribute Trap: Systematically Suboptimal Clarifying Questions After Tool-Call Retrieval▼4.3k2.00
[PDF]When Words Change but Meaning Remains: Robust Semantic-Stylistic Watermarking for LLMs▼1.2k1.67
[PDF]Increasing Width Allows Greedy Layer-wise Training to Rival End-to-End Backpropagation in Self-Supervised Learning▼3.7k2.25
[PDF]QUASAR: Lowering the Loss Floor of Quantization-Aware Training with Loss-Aware Reconstruction▲5093.00
[PDF]Does Quantization Preserve How Language Models Answer? Causal-Equivariant Quantization for Symbol-Binding Circuits▼4.5k2.25
[PDF]ComposeLong: Self-Supervised Composition of Long-Context Training Data from Short Texts▼5.9k2.33
[PDF]What Makes Agents Effective for Scientific Design? A Controlled Evaluation with Scientific Foundation Model Tools▼5.9k2.33
[PDF]Memorize, Adapt, Ignore: Diagnosing Robot Learning Mechanisms under Training Data Variation▼5.9k2.33
[PDF]VANTAGE: Temporally Structured Predictive Uncertainty for Probabilistic Egocentric Gaze Anticipation▼4.5k2.25
Withdrawn or desk rejected (1,836)
[PDF]Everything in Moderation: Interior Per-Domain Coverage Optima and Gap-Preserving Fixed-Budget Alignment in Multi-Domain Mid-TrainingDesk rejected
[PDF]When Is a Policy Violation a Model Failure? Policy Observability and Failure Attribution in Agent-Safety EvaluationDesk rejected