Reasoning behavior classifiers (Anthropic)
Only in system cards
Classifiers scan reasoning text from training episodes for deception, harmful actions and distress. Score is the share of outputs flagged.
In-house Anthropic monitoring of training transcripts. Results are given per category with no single overall rate.
Made by Anthropic Deception and alignment Tasks kept private Last in a system card May 2025
Last model added May 2025
- Level
- Expert Specialists with years of training, often over hours of work
- Scored
- Graded by a model A language model scores each response, usually against a rubric.
- Answers
- Open-ended Free-form text, documents, designs or media judged as a whole.
2
System cards report it, from 1 lab
No scores here yet
No lab, leaderboard or Epoch AI run has reported a score for Reasoning behavior classifiers (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.
2 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 |
|---|---|---|---|---|
| 22 May 2025 | Claude Opus 4 system card | Reasoning behavior classifiers on training outputs, no single overall. Deception overall 0.15%, harmful actions overall 0.00%, distress overall 0.00% (Table 4.1.6.A) | Alignment | |
| 24 Feb 2025 | Claude 3.7 Sonnet system card | Monitoring for concerning thought processes across 9,833 conversations, no single overall. Deception overall 0.37%, harmful actions overall 0.04%, distress overall 0% (Table 4) | Alignment |
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%