AI model builders mostly highlight unique benchmarks that act as flexible narrative tools for market positioning rather than standardized scientific measurements.
hub
21 NVIDIA Corporation
33 Pith papers cite this work, alongside 23 external citations. Polarity classification is still indexing.
33
Pith papers citing it
23
external citations · external index
hub tools
citation-role summary
background 3
citation-polarity summary
roles
background 3polarities
background 3representative citing papers
OVIG introduces an optimistic gradient-based verification framework for outsourced AI post-training that uses stride-sampled interval checks against an honest-replay boundary to achieve 0% attack success rate with low overhead.
SlotGCG uses Vulnerable Slot Score (VSS) to identify and target the most vulnerable prompt positions for adversarial token insertion, delivering 14% higher ASR than standard GCG and 42% higher against defenses.
Thermally activated clothing with thermochromic dyes and heaters creates dynamic adversarial patterns that evade AI surveillance in visible and infrared modalities while appearing ordinary when inactive.
WattLayer is a layer-wise energy estimation model achieving 19.6% median error on over 100k layers from 295 architectures across 3 tasks and 3 platforms, with generalization to new tasks via shared layers.
The paper demonstrates a black-box model extraction attack on graph classification models that leverages binary subgraph explanations to guide Monte Carlo edge sensitivity estimation with concentration guarantees.
A telemetry-driven cross-site router lets LLM inference run at wind-powered sites, cutting P99 latency by 22-52% over a strong baseline in a 64-GPU emulation.
A repeatable worksheet and human-reviewed expansion process turns expert-elicited AI use cases into 107 grounded scenarios to support consistent human-centered evaluations.
Scepsy schedules arbitrary multi-LLM agentic workflows on GPU clusters by constructing Aggregate LLM Pipelines from stable per-LLM execution time shares, then searching fractional GPU allocations, tensor parallelism, and replica counts to achieve up to 2.4x higher throughput and 27x lower latency.
New techniques for error-independent unified path variation, non-degenerate batched sampling, and flexible contraction accelerate tensor network quantum trajectory simulations by more than 10^8 times.
LLMs hallucinate citations at rates from 14.23% to 94.93%, with 1.07% of papers containing invalid citations and an 80.9% increase in 2025.
A Dirichlet-prior Bayesian estimator for model success probability replaces Pass@k, delivering faster-converging and more stable rankings with credible intervals on math benchmarks.
An accurate and trusted AI system cannot achieve human-level reasoning because there exist tasks easily solvable by humans but not by the system.
CDRO framework for grid-interactive cooling that adapts ambiguity sets, derives tractable reformulations, and shows near-zero violations plus 13.7pp cost reduction vs. Min-Max MPC in EnergyPlus simulations.
Experiment finds anthropomorphic language in AI descriptions does not substantially alter participants' perceptions compared to non-anthropomorphic versions.
Empirical analysis of 1,524 AI incident reports shows 83% arise from worker-AI trait misalignments, with 74% of those traceable to developers prioritizing efficiency over precision or personalization.
Filtering job posting data before LLM-assisted clustering and hierarchical labeling yields taxonomies with better AI skill coverage than unfiltered approaches.
A civilization's total cognitive computing capacity equals its power times compute efficiency, and at Type I scale this yields roughly one brain-equivalent per person.
Recruiters perceive themselves as retaining agency over GenAI in hiring pipelines, yet GenAI invisibly architects core evaluation inputs, producing only marginal efficiency gains at the cost of deskilling.
A survey of 232 papers frames the 'last mile' challenge of deploying multilingual language models under edge hardware constraints for Global South communities.
OSC separates token-persistent outlier channels in activations into a compact high-precision tensor for dual-path 4-bit GEMM computation, limiting accuracy loss to roughly 1-2 points on Qwen3 models while delivering up to 1.78x speedup over W8A8 baselines.
Performance analysis of multi-node LLM inference identifies all-reduce bottlenecks and introduces NVRAR hierarchical all-reduce achieving 1.9-3.6x lower latency than NCCL and up to 1.72x end-to-end batch latency reduction for Llama 3.1 405B in decode-heavy tensor-parallel workloads.
BaseRT achieves up to 1.56x higher LLM decode throughput than llama.cpp on Apple Silicon through native Metal kernel fusion and unified memory optimizations.
Federated training of Cox, DeepSurv, and Random Survival Forest models on cross-institutional breast cancer data outperforms local training and approaches centralized performance, with RSF showing the best overall balance.
citing papers explorer
-
Unsteady Metrics and Benchmarking Cultures of AI Model Builders
AI model builders mostly highlight unique benchmarks that act as flexible narrative tools for market positioning rather than standardized scientific measurements.
-
OVIG: Optimistic Verification of AI Training Integrity via Gradient Signals
OVIG introduces an optimistic gradient-based verification framework for outsourced AI post-training that uses stride-sampled interval checks against an honest-replay boundary to achieve 0% attack success rate with low overhead.
-
SlotGCG: Exploiting the Positional Vulnerability in LLMs for Jailbreak Attacks
SlotGCG uses Vulnerable Slot Score (VSS) to identify and target the most vulnerable prompt positions for adversarial token insertion, delivering 14% higher ASR than standard GCG and 42% higher against defenses.
-
Thermally Activated Dual-Modal Adversarial Clothing against AI Surveillance Systems
Thermally activated clothing with thermochromic dyes and heaters creates dynamic adversarial patterns that evade AI surveillance in visible and infrared modalities while appearing ordinary when inactive.
-
WattLayer: Get Layers Right to Estimate Inference Energy of Neural Networks
WattLayer is a layer-wise energy estimation model achieving 19.6% median error on over 100k layers from 295 architectures across 3 tasks and 3 platforms, with generalization to new tasks via shared layers.
-
Can Subgraph Explanations Be Weaponized to Steal Graph Neural Networks?
The paper demonstrates a black-box model extraction attack on graph classification models that leverages binary subgraph explanations to guide Monte Carlo edge sensitivity estimation with concentration guarantees.
-
CWind: A Cross-site Router for Large Language Model Inference Serving at Renewable Energy Farms
A telemetry-driven cross-site router lets LLM inference run at wind-powered sites, cutting P99 latency by 22-52% over a strong baseline in a 64-GPU emulation.
-
Towards Apples to Apples for AI Evaluations: From Real-World Use Cases to Evaluation Scenarios
A repeatable worksheet and human-reviewed expansion process turns expert-elicited AI use cases into 107 grounded scenarios to support consistent human-centered evaluations.
-
Scepsy: Serving Agentic Workflows Using Aggregate LLM Pipelines
Scepsy schedules arbitrary multi-LLM agentic workflows on GPU clusters by constructing Aggregate LLM Pipelines from stable per-LLM execution time shares, then searching fractional GPU allocations, tensor parallelism, and replica counts to achieve up to 2.4x higher throughput and 27x lower latency.
-
Accelerating Quantum Tensor Network Simulations with Unified Path Variations and Non-Degenerate Batched Sampling
New techniques for error-independent unified path variation, non-degenerate batched sampling, and flexible contraction accelerate tensor network quantum trajectory simulations by more than 10^8 times.
-
GhostCite: A Large-Scale Analysis of Citation Validity in the Age of Large Language Models
LLMs hallucinate citations at rates from 14.23% to 94.93%, with 1.07% of papers containing invalid citations and an 80.9% increase in 2025.
-
Don't Pass@k: A Bayesian Framework for Large Language Model Evaluation
A Dirichlet-prior Bayesian estimator for model success probability replaces Pass@k, delivering faster-converging and more stable rankings with credible intervals on math benchmarks.
-
Limitations on Accurate, Trusted, Human-level Reasoning
An accurate and trusted AI system cannot achieve human-level reasoning because there exist tasks easily solvable by humans but not by the system.
-
Grid-Interactive Thermal Management of AI Data Centers via Contextual Distributionally Robust Optimization
CDRO framework for grid-interactive cooling that adapts ambiguity sets, derives tractable reformulations, and shows near-zero violations plus 13.7pp cost reduction vs. Min-Max MPC in EnergyPlus simulations.
-
How Anthropomorphic Language Impacts Public Perceptions of AI
Experiment finds anthropomorphic language in AI descriptions does not substantially alter participants' perceptions compared to non-anthropomorphic versions.
-
The Quiet Path from Seemingly Minor Design Errors to Workplace AI Incidents
Empirical analysis of 1,524 AI incident reports shows 83% arise from worker-AI trait misalignments, with 74% of those traceable to developers prioritizing efficiency over precision or personalization.
-
Building a Custom Taxonomy of AI Skills and Tasks from the Ground Up with Job Postings
Filtering job posting data before LLM-assisted clustering and hierarchical labeling yields taxonomies with better AI skill coverage than unfiltered approaches.
-
The Cognitive Kardashev Scale: Quantifying the Material Envelope of Civilisational Computation
A civilization's total cognitive computing capacity equals its power times compute efficiency, and at Type I scale this yields roughly one brain-equivalent per person.
-
Resume-ing Control: (Mis)Perceptions of Agency Around GenAI Use in Recruiting Workflows
Recruiters perceive themselves as retaining agency over GenAI in hiring pipelines, yet GenAI invisibly architects core evaluation inputs, producing only marginal efficiency gains at the cost of deskilling.
-
Multilinguality at the Edge: Developing Language Models for the Global South
A survey of 232 papers frames the 'last mile' challenge of deploying multilingual language models under edge hardware constraints for Global South communities.
-
OSC: Hardware Efficient W4A4 Quantization via Outlier Separation in Channel Dimension
OSC separates token-persistent outlier channels in activations into a compact high-precision tensor for dual-path 4-bit GEMM computation, limiting accuracy loss to roughly 1-2 points on Qwen3 models while delivering up to 1.78x speedup over W8A8 baselines.
-
Understanding and Improving Communication Performance in Multi-node LLM Inference
Performance analysis of multi-node LLM inference identifies all-reduce bottlenecks and introduces NVRAR hierarchical all-reduce achieving 1.9-3.6x lower latency than NCCL and up to 1.72x end-to-end batch latency reduction for Llama 3.1 405B in decode-heavy tensor-parallel workloads.
-
BaseRT: Best-in-Class LLM Inference on Apple Silicon via Native Metal
BaseRT achieves up to 1.56x higher LLM decode throughput than llama.cpp on Apple Silicon through native Metal kernel fusion and unified memory optimizations.
-
Federated Survival Analysis in Healthcare: A Multi-Model Evaluation on Cross-Institutional Heterogeneous Breast Cancer Data
Federated training of Cox, DeepSurv, and Random Survival Forest models on cross-institutional breast cancer data outperforms local training and approaches centralized performance, with RSF showing the best overall balance.
-
Prioritization of Risks from Artificial Intelligence: A Delphi Study of 272 International Experts
Delphi study of 272 experts finds 18 of 24 AI risks >10% likely to cause catastrophe by 2030 in business-as-usual, dropping to five under mitigations; users and public most vulnerable, developers and governments most responsible.
-
Brief Announcement: Generative Markov Model for Distributed Computing Systems
A factorized generative Markov model is proposed for distributed computing systems to enable tractable simulation, inference, and policy learning, shown in a collaborative AI inference case study.
-
Human-aligned AI Model Cards with Weighted Hierarchy Architecture
Introduces CRAI-MCF, an eight-module framework distilling 217 parameters from 240 projects into a quantitative sufficiency criterion for cross-model LLM comparison grounded in Value Sensitive Design.
-
Taxing Artificial Intelligence
Taxation of AI activities can correct externalities, redistribute costs and gains, and support regulation, though instruments like corporate taxes, consumption taxes, and excises vary in feasibility, measurement challenges, and effects on innovation.
-
Europe and the Geopolitics of AGI: The Need for a Preparedness Plan
AGI may arrive by 2030-2040 and reshape global power balances, requiring Europe to close gaps in compute, talent retention, industrial adoption, and unified policy responses through a coordinated preparedness agenda.
-
Scrapyard AI
Obsolete AI models left behind by rapid development can be repurposed like scrap materials to analyze and communicate the environmental and social effects of global mining.
-
Continued AI Scaling Requires Repeated Efficiency Doublings
Continued AI scaling remains feasible only if efficiency doublings recur repeatedly to keep logical compute affordable.
-
Synthetic Reflections on Resource Extraction
A bespoke Urban Dwelling and Mining Index is introduced to improve multimodal AI models' assessment of mining operations' spatial distribution from satellite imagery.
- A Tool to Map AI Programs in the U.S.: A Snapshot from April 2026 and an Analysis of Requirements for AI Majors and Minors