02 · See it in action
Follow the steps
- Write the specific capability being claimed.
- Ask which tasks and environments were evaluated.
- Distinguish performance, breadth and degree of autonomy.
- Look for independent testing, failure cases and evaluation limitations.
Worked example
Claim: "This system solves coding puzzles."
Evidence needed: task set, scoring method, unseen examples, failures.
Unsupported leap: "Therefore it can reliably do every job."Read the example once, then explain each part aloud in your own words. If it includes Python, you can understand the idea first and run the code later in a local Python environment. Code is shown as text; this page does not execute it.
Sign in to save your reading progress.
Sign in