Skip to content
Apart Research
Apart Research
Contact us
Sprint projectAug 16, 2026Peshawar

Testing Whether an Affective-Empathy-Attenuated Persona Decouples a Functional Welfare Representation from Model Behavior

Asma Ahmed

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

Read the report

Report: Testing Whether an Affective-Empathy-Attenuated Persona Decouples a Functional Welfare Representation from Model Behavior

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

Labs use model behavior as a safety signal, but nobody has tested whether that signal survives a change in the model's disposition.

I built a real pipeline to test it: fine-tuned a persona with attenuated affective concern, causally validated a candidate internal axis via steering and ablation, and ran a real pressure battery, all on trained weights.

My pre-registered confirmatory test couldn't be computed at this scale. I report that as uncomputable, not null. Exploratory results show near-identical activation separation across conditions, which a six-step training run can't yet let me interpret either way.

The contribution is the pipeline, not a verdict, plus a reporting standard: disclose what couldn't be tested as rigorously as what was.

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 question is well-posed and the "uncomputable ≠ null" convention is a defensible norm to argue for. But six optimizer steps and 8–10 transcripts per arm means nothing here discriminates between the hypothesis being false and the intervention never happening

Cite this project

@misc{ahmed2026testing,
  title = {{Testing Whether an Affective-Empathy-Attenuated Persona Decouples a Functional Welfare Representation from Model Behavior}},
  author = {Asma Ahmed},
  year = {2026},
  month = aug,
  note = {Submitted to Digital Minds Research Sprint, an Apart Research Sprint},
  howpublished = {\url{https://apartresearch.com/sprints/projects/testing-whether-an-affectiveempathyattenuated-persona-decouples-a-functional-welfare-representation-from-model-behavior-1alk}},
  url = {https://apartresearch.com/sprints/projects/testing-whether-an-affectiveempathyattenuated-persona-decouples-a-functional-welfare-representation-from-model-behavior-1alk}
}

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