Evaluator Bench
Latent Space, 2026-06-22; retrieved 2026-09-15 tier 4 press (aggregator)confirmed

Red-Teaming after Mythos — Zico Kolter & Matt Fredrikson, Gray Swan

https://www.latent.space/p/gray-swan

Podcast/newsletter page; fetched 2026-09-15; span 'the adversarial red teaming tool that Anthropic used to evaluate the robustness of their models' present (Shade). Lead for X.1: a lab using the evaluator's product.

Signals citing this source

  • Gray Swan, X (against): A podcast interview reports Shade, Gray Swan's defense product, in use at Anthropic, a lab whose models Gray Swan red-teams.

Ledger rows citing this URL

  • none