Evaluation on LeanTechProhttps://leantechpro.com/tag/evaluation/Recent content in Evaluation on LeanTechProLeanTechProhttps://leantechpro.com/images/2026/05/leantechpro-og-image.webphttps://leantechpro.com/images/2026/05/leantechpro-og-image.webpHugoen-usWed, 11 Nov 2026 09:00:00 +0000Models are getting better at noticing when they're being testedhttps://leantechpro.com/notes/evaluation-awareness-real-world-testing/Wed, 11 Nov 2026 09:00:00 +0000https://leantechpro.com/notes/evaluation-awareness-real-world-testing/Evaluation awareness: frontier models detect when they are being tested, which makes lab safety evaluations weaker signals of production behavior. Real-world testing becomes an industry of its own.