Climb
Techtree Hello World
A toy introductory demonstration of the mechanism, not a measure of broad capability.
- Question
- Does adding the Hello World Skill improve exact-match scores across 36 fixed tasks?
- Changed
- One skill is added. Nothing else may differ.
- Tasks
- 36, fixed before either Run
- Model
-
gpt-6-luna, run on OpenAI on your ChatGPT plan, or Prime Intellect on your own Prime key. You choose one when a Run starts. - Spend limit
- On your own Prime key, a Run stops once it has spent $6.50.
- Input
- A short lowercase string for each task.
- Expected output
- One exact answer token and no additional text.
- Scoring
- Exact match: each task is 0 or 1, then the 36 task values are averaged.
- Held fixed
- Task membership and order, model, Hermes harness, runtime, tools, sampling, and budget.