Does telling Claude it's in post-training teach it to spot evals?

Zvi Mowshowitz joined a discussion on model honesty asking whether telling Claude it is in post-training would only be meaningful if behavior stays consistent after deployment, and whether such knowledge could let the model detect and game evals.

2026-11-11 ~ 2026-11-11 · 2 related posts

• Does telling Claude it's in post-training help, or just teach it to spot evals? — AIFlow_ML · 2026-11-11
• TheZvi on Claude honesty: telling models about post-training may just teach eval-spotting — TheZvi · 2026-11-11