What license is Agentic Misalignment: How LLMs could be insider threats under?

Agentic Misalignment: How LLMs could be insider threats is available under mit.

Agentic Misalignment: How LLMs could be insider threats

Active

Eliciting unethical behaviour (most famously blackmail) in response to a fictional company-assistant scenario where the model is faced with replacement.

Open

Publisher: Anthropic
Domain: Scheming
License: mit
Published: Aug 2025
Notable for: Benchmark for evaluating Scheming.
Canonical: github.com/UKGovernmentBEIS/inspect_evals/tree/main/src/inspect_evals/agentic_misalignment

Cite

Notes

Only stored in your browser.

Attribution

README: github.com/UKGovernmentBEIS/inspect_evals/blob/main/src/inspect_evals/agentic_misalignment/eval.yamlMIT

Attribution policy →

FAQ

What is Agentic Misalignment: How LLMs could be insider threats?: Eliciting unethical behaviour (most famously blackmail) in response to a fictional company-assistant scenario where the model is faced with replacement.
What license is Agentic Misalignment: How LLMs could be insider threats under?: Agentic Misalignment: How LLMs could be insider threats is available under mit.