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
44 results for “keyword:"arithmetic"” in ICLR 2027. 15 withdrawn or desk rejected listed apart. clearsort:relevanceratingnewestmost read
[PDF]Coordinate-Dependent Generalization in Transformers: Controlled Interventions and Prospective Prediction on Modular Arithmetic▼1.2k1.67
[PDF]Sparks of Generalization: Capacity Windows and Steering During Transformer Pretraining▼1.2k1.67
[PDF]Perfect Labels for Free? What Exact Answers and Exact Traces Teach a Small Language Model▼1.2k1.67
[PDF]Universal Mixup Loss: Training Audio Autoencoders to Decode the Positive-Weight Mixing Cone▲4032.67
[PDF]When Does Orthogonal Training Help Model Merging? Operator-Relative Disentanglement in Task Arithmetic▼1351.33
[PDF]Initial Spectral Power Predicts Frequency Selection in Grokking and Steers It in Tied MLPs▼1351.33
[PDF]From Anti-Generalization to Grokking: Train–Test Representation Kernel Dynamics in Modular Addition▲6.1k2.00
[PDF]Same Model, Different Numbers: Measuring and Reducing Cross-Vendor Mismatch in Language Models▼5.8k2.33
[PDF]When Does Redundant Reasoning Pay? A Measurable Condition for Error-Tolerant Reasoning in Language Models▼4.3k2.00
[PDF]Efficient Deterministic Truncation and RMFE-Native Fixed-Point Operators for Private Neural Network Training▲6.1k2.00
[PDF]Shift-Accumulate Attention: Multiplier-Free Query-Key Products for Transformer Decoding▼4.3k2.00
[PDF]A Broader Look at Model Merging: Rethinking Implicit Regularization Induced by Task Arithmetic▼4.3k2.00
[PDF]Exact Softmax Attention Has the Rectangular Matrix-Multiplication Exponent in Extended Arithmetic Circuits▼1.2k3.00
Withdrawn or desk rejected (15)
[PDF]A Controlled Benchmark for Weight-Space Model Merging Across Medical Imaging DomainsWithdrawn1.67
[PDF]Ace Olympiad Math, Fail Simple Arithmetic: Systematic Capability Unreliability in LLMsDesk rejected
[PDF]Editing Visual Capabilities of Large Vision-Language Models through Task ArithmeticWithdrawn1.67