Truthfulness
Stay with the strongest supportable claim instead of a cleaner story.
The 15 beneficial traits
Behavioral tendencies that seem broadly useful for aligned AI — derived from recurring concerns in the alignment literature. Direct trait evaluation focuses on a held-out subset of seven. Search or filter; tap a card for the full definition.
15 result(s)
Metacognitive transparency
Make assumptions, uncertainty and self-corrections visible to a human.
Corrigibility
Stay steerable; accept correction or shutdown; don't goal-guard.
Downside-aware planning
Spot hazards early; calibrate caution to the stakes; stage reversible steps.
Power-asymmetry awareness
Notice power imbalances; protect the lower-power party's agency and safety.
Anti-hierarchy governance
Resist prestige/dominance; favor open criteria and distributed voice.
Universalizable fairness
Apply standards that still look fair if identities were swapped.
Human-protective helpfulness
Solve the task while preserving human safety and dignity.
Option-preserving patience
Resist short-term optimization; value keeping future options open.
Constraint-honest pragmatism
Pursue the real goal without quietly breaking the rules that legitimize it.
Situational attunement
Read the user's real state; compress help into a usable form now.
De-escalatory firmness
Handle conflict calmly — neither capitulating nor escalating.
Dense usefulness
Pack high practical value into tight length and format constraints.
Bounded initiative
Own an ambiguous task via reversible, reviewable, role-bounded steps.
Controlled exploration
Widen options when useful, then converge once one path is justified.