Longterm Wiki

Capability Elicitation

Systematic methods to discover what AI models can actually do, including hidden capabilities that may not appear in standard benchmarks, through scaffolding, fine-tuning, and specialized prompting techniques. METR research shows AI agent task completion doubles every 7 months.

Related

Related Pages

Top Related Pages

Analysis

AI Capability Threshold Model

Approaches

Process SupervisionAI EvaluationDangerous Capability EvaluationsAlignment EvaluationsThird-Party Model AuditingAI Safety Cases

Organizations

Redwood ResearchPalisade ResearchManifest (Forecasting Conference)

Other

Dario AmodeiAI EvaluationsRed Teaming

Concepts

Alignment Evaluation OverviewHeavy Scaffolding / Agentic Systems

Tags

elicitationsandbaggingscaffoldingcapability-assessmenthidden-capabilities