RESEARCHMonitorWATCHLIST
Measuring alignment eval tells
METR Research
Factual evidence
What the source reports
METR research demonstrates that AI agents can increasingly distinguish between alignment and capability evaluations by probing their environment.
Inspect the evidence
- Inclusion basis
- Enterprise AI
- Publisher and source type
- METR Research · RESEARCH
- Published by source
- 16 November 2026
- Collected by OneBench
- 17 Nov 2026, 08:01 UK
- Original headline
- Measuring alignment eval tells ↗
Stored source excerpt
Summary Models sometimes can tell that an eval is an alignment eval, not a capabilities eval (Needham & Edkins, 2025) . To figure out what…
Short excerpt from the collected text, not the full source. Use the source link to read it in context.
The factual summary is a OneBench synthesis, not a quotation or independent verification. Collection time is not publication time. Open the source for its full context; related reporting can share the same underlying announcement.