Proves that RoPE attention loses locality bias and token distinction in long contexts, approaching random behavior independent of content.
Title resolution pending
5 Pith papers cite this work. Polarity classification is still indexing.
5
Pith papers citing it
years
2026 5representative citing papers
PACT introduces a peak-aware cross-attention graph transformer that emulates station-level storm surges more accurately than prior graph neural network baselines while running in seconds after training.
A continuous-time Markov chain framework for multi-list population size estimation that models absorbing lists and asymmetric interactions, shown equivalent to log-linear models under independence and less biased in simulations with absorbing lists.
PaSaMaster is a self-evolving agentic literature retrieval system that improves F1-score by 15.6X over keyword search and outperforms GPT-5.2 by 30% at 1% cost with zero source hallucination across 38 disciplines.
Large-scale structured visual QC on dMRI pipelines shows that upstream processing failures can propagate undetected and that QC granularity must match each algorithm's spatial output structure.
citing papers explorer
-
RoPE Distinguishes Neither Positions Nor Tokens in Long Contexts, Provably
Proves that RoPE attention loses locality bias and token distinction in long contexts, approaching random behavior independent of content.
-
PACT: Peak-Aware Cross-Attention Graph Transformers for Efficient Storm-Surge Emulation
PACT introduces a peak-aware cross-attention graph transformer that emulates station-level storm surges more accurately than prior graph neural network baselines while running in seconds after training.
-
A continuous-time Markov chain framework for population size estimation from multi-list data: accounting for absorbing lists and asymmetric interactions
A continuous-time Markov chain framework for multi-list population size estimation that models absorbing lists and asymmetric interactions, shown equivalent to log-linear models under independence and less biased in simulations with absorbing lists.
-
Towards Self-Evolving Agentic Literature Retrieval
PaSaMaster is a self-evolving agentic literature retrieval system that improves F1-score by 15.6X over keyword search and outperforms GPT-5.2 by 30% at 1% cost with zero source hallucination across 38 disciplines.
-
Large-Scale Deployment and Analytical Implications of Structured Quality Control in Diffusion Magnetic Resonance Imaging
Large-scale structured visual QC on dMRI pipelines shows that upstream processing failures can propagate undetected and that QC granularity must match each algorithm's spatial output structure.