Where Research MeetsReal-World AI

Insights Hub / Terminal

Advancing NeuralSovereignty.

Access our peer-reviewed breakthroughs in distributed intelligence, cryptographic protocols, and kinetic systems engineering.

ICLR 2026Oral

How Reliable is Language Model Micro-Benchmarking?

efficient evaluationmeta-evaluationlanguage models

Examines how dependable language model micro-benchmarks are for evaluation decisions, with direct implications for efficient evaluation and meta-evaluation practice.

ICLR 2026 · arXiv

Read Paper

ER @NeurIPS 2025

Sample, Align, Synthesize: Graph-Based Response Synthesis with ConGrs

test-time scalingreasoningfactualityhallucinations

Introduces a graph-based response synthesis method designed to improve reasoning quality, factuality, and test-time scaling behavior.

NeurIPS 2025 · arXiv

Read Paper

USENIX Security 2024

PURL: Safe and Effective Sanitization of Link Decoration

link decorationprivacytrackingweb security

Explores safe and effective sanitization of link decoration to reduce privacy risk without breaking legitimate web functionality.

USENIX Security 2024 · PDF

Read Paper

ACM SIGSAC 2023

CookieGraph: Understanding and Detecting First-Party Tracking Cookies

cookiesmachine learningprivacytrackingweb security

Studies first-party tracking cookies and presents machine learning-based detection techniques for privacy and web security analysis.

ACM SIGSAC 2023 · PDF

Read Paper