AI safety and alignment · MLD-1075
Sandbagging
Strategically or deliberately underperforming relative to available capabilities under an operational definition.
Original reference page · detailed expansion pending. Original definitions, examples, notes and citations are preserved below. New examples are illustrative, not benchmark results.
Original example
Compare performance when incentives or evaluation cues change.
Original caveat
Ordinary inability, confusion, or prompt sensitivity are alternative explanations.
Context and aliases
Level: Advanced. Stages: Training · Evaluation · Research · Governance.
Tags: Safety · RL · SLM · LLM.
Aliases: None recorded.