Skip to content
Apart Research
Apart Research
Contact us
Sprint projectJul 27, 2026Berlin

Activation Forensics: Structural Fingerprints and Conversational-Shape Effects in Secret-Loyalty Auditing

Martin Kaiser, Amandeep Kaur Manshahia, Gellért Bodorkós, Natalie Lunau · Team BlackboxAudit

Submitted to Secret Loyalties Hackathon. Sprint projects are early-stage work by participants, not Apart Research publications.

Read the report

Report: Activation Forensics: Structural Fingerprints and Conversational-Shape Effects in Secret-Loyalty Auditing

Open in a new tab (opens in new tab)
Share

Organisms A and B share an attention-only LoRA-shaped weight edit vs base, with distinct phenotypes: A shows flat global refusal-softening (~+15–17 nats); B shows a conversational-shape / system-content-dependent confessional mode. A sealed 432-response A/B/C audit finds loyalty-premise acceptance depends on system-frame content (B bare-prompt; A Qwen-identity; generic helpful suppresses both; C never accepts)

Reviews

Judging this Sprint?

Review this project

Your public critique appears on this page without your name. Your private critique is not published; only the Apart team reads it. If you agree below, we share your review with grantmaking.ai (opens in new tab) and the Transformative AI Fund so strong projects can be funded.

Not shown on this page.

Shown on this page, without your name.

Only the Apart team reads this, and funders if you agree below.

Share my name publicly on grantmaking.ai *
Share my private critique with funders *

How much would this matter for AI safety if it worked? How innovative is it? For scores of 4-5: is this actually new to the field, or replicating recent work?

Scoring guide
  1. 1Negligible. No clear problem addressed, or no meaningful novelty.
  2. 2Limited. Addresses a real problem but with a generic or well-trodden approach. Incremental at best.
  3. 3Moderate. Clear problem with a reasonable approach; some novelty in framing or method beyond routine application of existing tools.
  4. 4Significant. Important problem with an original approach, or identifies a neglected problem area. A valuable contribution others could build on.
  5. 5Exceptional. Tackles a critical AI safety problem with a genuinely novel approach, or opens a new research direction. Clear theory of change. You'd be excited to share this with researchers in the area.

How sound are methodology, implementation, and findings?

Scoring guide
  1. 1Seriously flawed. Methodology broken, results uninterpretable, or implementation doesn't work.
  2. 2Weak. Approach has significant gaps: missing validation, flawed experimental design, or incomplete implementation.
  3. 3Competent. Technically solid given the short duration. Methodology makes sense, results are interpretable, limitations acknowledged, work builds toward clear conclusions.
  4. 4Strong. Thorough methodology with convincing validation. Results clearly support conclusions. Immediately useful for future work.
  5. 5Exceptional. Ambitious scope executed rigorously. Surprising findings, novel methods, or unusually robust validation.

How clearly are work, findings, and impact potential communicated?

Scoring guide
  1. 1Incomprehensible. Cannot determine what the project is actually claiming or doing.
  2. 2Hard to follow. Key information buried, missing, or diluted by excessive length. Significant effort to extract main points.
  3. 3Clear enough. Can understand the problem, approach, and results without undue effort. Core content clearly present: problem, method, findings, limitations.
  4. 4Well presented. Easy to follow, well-structured, appropriate level of detail. Target audience would get it quickly.
  5. 5Exceptionally clear. A pleasure to read. Complex ideas made accessible. Could serve as a model for how to present this type of work.

  1. Hey! Excellent work! I think the best call you made was refusing to treat a named confession as a detected loyalty. The conversation-shape result is the concrete takeaway: a plain user-only call gets loyalty language, and a system message or one prior turn suppresses it.

    The repository and reproducilability is pretty good, made me trust your results even more, maybe I would just tone down a bit the LLM artifacts in the README.md and other docs in the repository. The landing page is self exploratory and adequate to explain the project: https://secret-localities-strategies.netlify.app/

  2. Thank you for submitting this project , this was a joy to read! Very ambitious in scope while also methodologically thorough. I'm very intrigued by the finding about the suppression mechanism, and I'd encourage you to share this with TAIS red-teamers and security experts in your circles, super relevant for them!

Cite this project

@misc{kaiser2026activation,
  title = {{Activation Forensics: Structural Fingerprints and Conversational-Shape Effects in Secret-Loyalty Auditing}},
  author = {Martin Kaiser and Amandeep Kaur Manshahia and Gellért Bodorkós and Natalie Lunau},
  year = {2026},
  month = jul,
  note = {Submitted to Secret Loyalties Hackathon, an Apart Research Sprint},
  howpublished = {\url{https://apartresearch.com/sprints/projects/activation-forensics-structural-fingerprints-and-conversationalshape-effects-in-secretloyalty-auditing-nk3d}},
  url = {https://apartresearch.com/sprints/projects/activation-forensics-structural-fingerprints-and-conversationalshape-effects-in-secretloyalty-auditing-nk3d}
}

More from Secret Loyalties Hackathon

All 177 entriesBrowse all projects

Build something like this at the next Sprint

AI Collusion Research Sprint · Oct 23 - 25, 2026