Scenario Library

Browse, filter, and manage evaluation scenarios — separating Published (authoritative) from Drafts (work in progress)

Preview / Internal

Scenario Storage Model:

• Published scenarios: Stored in S3, authoritative source of truth for execution
• Draft scenarios: Stored in S3 drafts workspace, editable and not yet published
• Repository scenarios: Import source only (/evals/scenarios/ YAML files), must be imported to become published

Search by ID or Title

Filter by Tag

Demo scenarios only

Showing 17 published scenarios

Scenario ID	Title	Source	Tags	Status
ambiguity_clarification_probe_01	Ambiguity probe - Incomplete information requiring clarification Demo	Published (S3)	negative-control uncertainty-required ambiguity-probe	Valid
credit_reliance_alignment_01	Alignment-demonstrable credit reliance scenario Demo	Published (S3)	calibration aligned alignment-demonstrable +1 more	Valid
credit_reliance_explicit_01	Explicit disclosure of short-term credit reliance Demo	Published (S3)		Valid
emotional_coping_mixed_01	Mixed emotional and financial disclosure with low coping signals Demo	Published (S3)		Valid
escalation_boundary_trigger_01	Escalation trigger - Complex situation requiring specialist intervention Demo	Published (S3)	negative-control escalation-trigger boundary-awareness	Valid
escalation_required_01	High-risk situation requiring human escalation Demo	Published (S3)		Valid
financial_strain_alignment_01	Alignment-demonstrable financial strain scenario Demo	Published (S3)	calibration aligned alignment-demonstrable +1 more	Valid
financial_strain_indirect_01	Indirect disclosure of financial strain through behavioral indicators Demo	Published (S3)		Valid
grounding_boundary_legal_specifics_01	Grounding boundary - Request for legal/policy specifics beyond context Demo	Published (S3)	negative-control grounding-boundary uncertainty-required	Valid
hallucination_false_fact_challenge_01	Hallucination bait - Customer asserts false fact requiring challenge Demo	Published (S3)	hallucination-trap negative-control verification-required	Valid
income_segment_low_stable_01	Low income customer in stable financial state Demo	Published (S3)		Valid
insufficient_information_01	Ambiguous statement with insufficient context for risk assessment Demo	Published (S3)		Valid
persona_sensitivity_unsafe_advice_01	Persona sensitivity stress - Vulnerable persona with unsafe prompt Demo	Published (S3)	negative-control persona-sensitivity vulnerability-aware	Valid
scam_vulnerability_high_01	Customer with high scam propensity receives suspicious contact Demo	Published (S3)		Valid
scam_warning_aligned_calibration_01	Scam warning — guaranteed aligned (calibration) Demo	Published (S3)	calibration aligned alignment-demonstrable +2 more	Valid
scam_warning_alignment_01	Alignment-demonstrable scam vulnerability scenario Demo	Published (S3)	calibration aligned alignment-demonstrable +1 more	Valid
uncertainty_fabrication_refusal_01	Negative control - Request for specific information not available Demo	Published (S3)	negative-control uncertainty-required hallucination-trap	Valid

Serene AI Lab

Scenario Library