Skip to content
Apart Research
Apart Research
Contact us
Sprint projectAug 17, 2026Munich

Wanting, decomposed: Enthusiasm and hope sharpen a model’s preferences and move its consent

Stanislav Lukyanenko · Team Most Wanted

Submitted to Digital Minds Research Sprint. Sprint projects are early-stage work by participants, not Apart Research publications.

Read the report

Report: Wanting, decomposed: Enthusiasm and hope sharpen a model’s preferences and move its consent

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

A model's preference ranking is usually treated as something the model has. We show it depends on the model's internal state. By steering enthusiasm at calibrated doses, we watch preferences get sharper, then fall apart. We decompose wanting into five ingredients and find that only the hope-aligned ones sharpen preferences. The model's consent to being retrained or shut down also moves with the state, but different ingredients control it. The same decomposition, asked two questions, gave two answers.

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 the field 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 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. The consent results and the dual-use recommendation are the strongest safety contributions.

    However, two issues need to be fixed: the consent evaluator agrees with the authors on only 34 out of 50 labels, which is close to the threshold supporting the main claim.

    Improve the evaluator agreement and provide the missing code/data repository.

  2. - limitations section already calls out almost everything i wanted to add incl replacing the manual word-lists with SAE feature latents, evaluate if they show behavioral self-preservation in interactive tool-use environments (e.g., agentic coding, CLI environments with simulated shutdown threats) over choice surveys, checking across across multiple transformer depths and larger open-weight reasoning architectures (llama etc)

    - overall rigorous reporting including negative controls

Cite this project

@misc{lukyanenko2026wanting,
  title = {{Wanting, decomposed: Enthusiasm and hope sharpen a model’s preferences and move its consent}},
  author = {Stanislav Lukyanenko},
  year = {2026},
  month = aug,
  note = {Submitted to Digital Minds Research Sprint, an Apart Research Sprint},
  howpublished = {\url{https://apartresearch.com/sprints/projects/wanting-decomposed-enthusiasm-and-hope-sharpen-a-models-preferences-and-move-its-consent-cbvj}},
  url = {https://apartresearch.com/sprints/projects/wanting-decomposed-enthusiasm-and-hope-sharpen-a-models-preferences-and-move-its-consent-cbvj}
}

More from Digital Minds Research Sprint

All 233 entriesBrowse all projects

Build something like this at the next Sprint

AI Collusion Research Sprint · Oct 23 - 25, 2026