Looped LLMs converge to distinct cyclic fixed points per layer, repeating feedforward-style inference stages across recurrences.
On vanishing gradients, over- smoothing, and over-squashing in gnns: Bridging recurrent and graph learning.arXiv preprint arXiv:2502.10818
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.LG 3verdicts
UNVERDICTED 3roles
background 1polarities
background 1representative citing papers
HISTOGRAPH applies unified layer-wise attention followed by node-wise attention over historical GNN activations to improve graph classification, especially in deep models.
Introduces quantitative error feedback from digital filter techniques to exactly compensate quantization noise in graph filtering, with closed-form optimal coefficients for deterministic, random-graph, and asynchronous scenarios.
citing papers explorer
-
A Mechanistic Analysis of Looped Reasoning Language Models
Looped LLMs converge to distinct cyclic fixed points per layer, repeating feedforward-style inference stages across recurrences.
-
Learning from Historical Activations in Graph Neural Networks
HISTOGRAPH applies unified layer-wise attention followed by node-wise attention over historical GNN activations to improve graph classification, especially in deep models.
-
Quantitative Error Feedback for Quantization Noise Reduction of Filtering over Graphs
Introduces quantitative error feedback from digital filter techniques to exactly compensate quantization noise in graph filtering, with closed-form optimal coefficients for deterministic, random-graph, and asynchronous scenarios.