
Held-Out Generation and Contamination Probes: Testing Agent Evals Against Memorization
pr0h0•
A hands-on look at why benchmark wins like Cloudflare's Clef-vs-Jev claim deserve scrutiny, paired with Google's anti-memorization technique and a practical checklist of held-out generation, dynamic task synthesis, and contamination probes for evals that can't be gamed.
evalsbenchmarkingai-agents