DEDUPED Hot AI repos, turned into use cases Get 5 a day

How to reproduce the identifier-fidelity probe experiment with Dynolab

Rerun the saved Qwen3.8 probe and compare it with the token-count baseline.

  • Research
  • Advanced
  • 2 hours
  • From the README
  • Any agent

Authorization notice. Dynolab is dual-use. Only use it on software, sites and systems you own or have explicit permission to test.

Who it is forAn interpretability researcher who wants to check a published probe result on their own Mac.

The problemA probe result is only useful if you can rerun it, see the held-out numbers and check the control baselines yourself.

The outcomeA local rerun of the saved layer-32 probe with held-out results and input-only controls, set against the README's reported figures.

What you need first

  • An Apple Silicon Mac running macOS 14 or later
  • The Dyno app downloaded from the latest GitHub release
  • An MLX model downloaded through Discover in the app

Steps

  1. Open examples/identifier-fidelity-probe/README.md in the repo.
  2. Start the Qwen3.8 model in Dyno Models, then follow the steps in that README.
  3. Watch the native Lab walkthrough at https://dynolab.dev/probe-example.html to see the expected layout.
  4. Compare your held-out results with the README's figures: 7 missing-reference outcomes caught with 2 false alarms, and the token-count baseline.

Prompt for your agent

In the canivel/dynolab repo, follow examples/identifier-fidelity-probe/README.md to run the saved probe on the model <model-name>. Expected output: held-out probe metrics next to the input-only control baselines. Verify that your numbers are reported as an exploratory robustness check, not as a validated safety monitor.

Caveat. The README says this is an exploratory robustness experiment, not a validated safety monitor, and its small held-out test is limited.

Share: X LinkedIn Hacker News Email Copy link

Disclaimer

Deduped lists open-source repositories that belong to other people. We do not own, maintain, audit, or vouch for any of them, and a repo appearing here is not an endorsement or a safety check. The use cases, steps, and prompts are written by AI and may be incomplete, outdated, or just wrong. Where a use case is marked "Seen in the wild", we link to public posts where someone described something similar, but we have not verified those posts and the use case itself is still our AI's own write-up. Use all of this at your own risk. Only run code, commands, or agents on systems and accounts you are authorized to use, and read the code and every command before you run it, especially anything that touches credentials, files, or the network. Nothing here is professional, security, legal, or financial advice. All trademarks and project names belong to their respective owners, and their use here does not imply affiliation. Content is provided as is, without warranties of any kind. Found a mistake, a safety problem, or something you want removed? Email [email protected] and we will look at it promptly. Full disclaimer page