Explore Prompts

Define a safe logging policy: what to log for debugging, what to redact, retention windows, access controls, and anonymization. Include examples and anti-examples.

Design a zero-trust approach: secrets management, short-lived credentials, environment segregation, and auditing. Include how to test that the AI cannot access restricted secrets.

Write explicit requirements to prevent deceptive behavior: transparency rules, auditability, and logging. Include tests to detect hidden policy violations or manipulation attempts.

Design a safe self-reflection procedure: bounded critique, avoiding infinite loops, and ensuring critique is grounded in evidence. Provide stop rules and performance safeguards.

Create a cross-validation pattern: primary model proposes, secondary model critiques/verifies, and a policy gate decides. Include how to handle disagreements and measure error reduction.

Design a multi-agent architecture with safety: role separation, constrained tools, message filtering, and cross-checks. Include failure modes and an evaluation plan.

Create a traceability matrix linking requirements/specs to tests and monitors. Provide a template and a worked example for a tool-using assistant system.

Describe a high-level approach to align reward signals with safe behavior: preference data guidelines, reward hacking risks, and validation. Keep it conceptual and focused on safety.

Create an annotation guide: definitions, examples, severity levels, and how to handle ambiguity. Include training exercises and a QA process for reviewer consistency.

Help me design a domain-specific safety benchmark: representative tasks, policy-sensitive cases, and adversarial cases. Include labeling guidelines and inter-annotator agreement checks.

Design drift detection: changes in user queries, outcome distributions, error types, and model behavior. Include thresholds and a playbook for when drift is detected.

Create a stress test plan: malformed inputs, long-context traps, conflicting instructions, and toxic content probes. Provide how to automate and score robustness over time.