Tracking frontier model releases, technical reports from leading labs, empirical safety evaluations, and open-source artifacts.
Direct primary sources on reasoning architectures, training mixture disclosures, and weights releases.
Rigorous benchmarks on latent misalignment, deceptive tendencies, and corrigibility assessments.
Deterministic compute bounds, inference budgeting, and edge denial-of-service failsafes.
As general-purpose models scale, rigorous evaluations are required to identify potential dual-use biological, cyber-offensive, situational awareness, and self-proliferation capabilities before deployment. We introduce an...
Polysemanticity—where individual neurons activate on multiple unrelated concepts—has long impeded mechanistic interpretability. We apply sparse autoencoders to intermediate activation states of a transformer language mod...
A corrigible artificial agent is one that tolerates or assists in its own modification, correction, or shutdown by its human operators. This paper investigates theoretical hurdles in designing utility functions that ince...
You are about to share this research paper context with your Ari companion. No personal browsing history or private memory is accessed.
Authenticated members bypass Turnstile challenges and gain personal library & discourse privileges.