neuron-explorer
Steering-first neuron explorer for Qwen3.5-4B: push every MLP neuron, keep every rollout, browse the results
Grades agent outputs against an eval set or a production outcome tag, then catches quality drift with a statistical significance check. The LLM judge gets priced too. Validated on a live Claude run: $
Developer toolsPython · ShellGitHub @TAIPANBOXOnlineOwner unconfirmed
Have you tried it? One tap is enough. No sign-in needed.
Owners of public repositories under a personal GitHub account can sign in to verify ownership and manage the introduction themselves. No administrator approval is needed.
Sign in, generate a verification file, and commit it to this repository’s default branch within 30 minutes. Files on other branches or forks do not count. Then check the file below to manage the project without administrator approval. Keep the file to retain your access.
If needed, you will go to sign-in. After signing in, return here and press your chosen verification button again.
Run /nomorevibe verify in your project folder to verify ownership and manage the introduction yourself.
Verify in one minuteAsk about other verification options
Quality evaluation and drift for AI agents: the quality plane of the agent-governance stack
Taken from the repository README
Loading the full README. The saved excerpt is shown below.
Grade an agent's outputs against an eval set or a production outcome tag, then catch when quality drifts.
Line coverage of the verdryx package: 97% (1510/1556 statements).
@measured .venv/bin/pytest --cov=verdryx --cov-report=term-missing 2026-09-25.
Unlike the tests badge above, this figure is not gated (see CLAUDE.md invariant
11): recomputing it inside a gate would mean running the full suite inside a
gate, which is what the other gates already run once per CI job, so it is
restated by hand at the point it is measured instead.Read the full README ↗The full README is unavailable right now. Read it at the original link. The saved excerpt is shown below.
Grade an agent's outputs against an eval set or a production outcome tag, then catch when quality drifts.
Line coverage of the verdryx package: 97% (1510/1556 statements).
@measured .venv/bin/pytest --cov=verdryx --cov-report=term-missing 2026-09-25.
Unlike the tests badge above, this figure is not gated (see CLAUDE.md invariant
11): recomputing it inside a gate would mean running the full suite inside a
gate, which is what the other gates already run once per CI job, so it is
restated by hand at the point it is measured instead.Read the full README ↗