Open benchmark for power-constrained LLM inference scheduling. Tokens per watt as a first-class metric.
-
Updated
Jun 26, 2026 - Python
Open benchmark for power-constrained LLM inference scheduling. Tokens per watt as a first-class metric.
A 10-day public experiment analyzing rule-level incentives, reward geometry, and security in the Bittensor protocol (AI + Crypto). Includes detailed daily analyses, first-principles reasoning, and diagnostic metrics.
Interactive verifier/optimizer lab for Self-Improving AI Goodhart failures
Add a description, image, and links to the goodhart topic page so that developers can more easily learn about it.
To associate your repository with the goodhart topic, visit your repo's landing page and select "manage topics."