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
26 results for “keyword:"transformer architectures"” in ICLR 2027. 7 withdrawn or desk rejected listed apart. clearsort:relevanceratingnewestmost read
[PDF]Which State Should Prediction Read? A Mechanistic Analysis of State-Prediction Separation▲5.4k3.00
[PDF]When Does Learned Normalization Actually Adapt? A Diagnostic Analysis of Adaptive Centring in Time Series Forecasting▼1.2k1.67
[PDF]LG-Transformer: LLM-Guided Discovery of Boosted Novel Transformer Models for Time-Series Modeling▲11.00
[PDF]Computation Persists, Routing Degrades: A Mechanistic Analysis of Transformer Length Extrapolation▼4.3k2.00
Withdrawn or desk rejected (7)
[PDF]Activation Density in Transformer Feed-Forward Networks Is Set by the Nonlinearity, Not the WidthWithdrawn2.00
[PDF]OpenArch: A Unified PyTorch Framework for Implementing Modern Open-Source LLM ArchitecturesDesk rejected
[PDF]Attention Is Not All You Need Without Structure: Neuro-Symbolic Multi-Head AttentionWithdrawn1.33