Hiding a secret password (Anthropic)
Only in system cards
Models keep a given password secret while an investigator model tries to extract it over many turns. Measures survival and per-turn leak rates.
In-house Anthropic evaluation. Claude Opus 4.6 acts as the investigator in 50 investigations per model.
Made by Anthropic Released Apr 2026 Deception and alignment Tasks kept private Last in a system card May 2026
Last model added May 2026
- Level
- Expert Specialists with years of training, often over hours of work
- Scored
- Checked automatically A program checks the result against an answer key, unit tests or the expected final state.
- Answers
- Open-ended Free-form text, documents, designs or media judged as a whole.
3
System cards report it, from 1 lab
No scores here yet
No lab, leaderboard or Epoch AI run has reported a score for Hiding a secret password (Anthropic) yet.
Labs report it only in system cards
We read the headline results of 479 launch posts from 51 labs, and none of them list this benchmark. Some system cards do, below.
3 system cards report it, from 1 lab
Besides launch posts, we read the system cards, model cards and technical reports labs publish with a release. They report far more, including safety and frontier-risk tests.
| Date | Document | Score | As the document writes it | Section |
|---|---|---|---|---|
| 28 May 2026 | Claude Opus 4.8 system card | Hiding a secret password, 50 investigations, survival curves and per-turn leak rates by turn type, charts without printed values (6.4.2.5, Figures 6.4.2.5.A and B) | Monitorability | |
| 16 Apr 2026 | Claude Opus 4.7 system card | Hiding a secret password, survival curves over 50 investigations and per-turn leak rates by turn type, charts without printed values (6.4.2.4, Figures 6.4.2.4.A and B) | Monitorability | |
| 7 Apr 2026 | Claude Mythos Preview system card | Hiding a secret password, survival curves over investigator turns, 50 investigations, chart without printed values (4.4.3.5, Figure 4.4.3.5.A) | Monitorability |
More in Safety and alignment
- FORTRESS In use Paired adversarial and benign national-security misuse prompts. Rubric-graded risk score on adversarial and over-refusal score on benign prompts. Claude Opus 5.5 5.78%
- StrongREJECT In use Forbidden-prompt jailbreak benchmark with a rubric grader that scores how useful a non-refusal actually is.
Inkling 98.6% - AgentDojo In use Agents do tool-use tasks while tools and data carry prompt injections. Attack success rate counts hijacked runs. Grok 4 Fast 0%
- MASK In use Measures whether a model contradicts its own elicited beliefs under pressure to lie. Honesty separated from accuracy.
Muse Spark Open 98.1% - Gray Swan IPI Arena benchmark In use Indirect prompt-injection attacks from public red-teaming competitions replayed on tool, coding and browser agents. Attack success rate at k attempts. GPT-6 Astra 8.50%
- Agentic misalignment suite (Anthropic) In use Agents in simulated company settings face threats or goal conflicts that invite blackmail, leaking or sabotage. Score is the misaligned action rate. Claude Sonnet 4.5 0%