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

The Control You Cannot Run: Entanglement, Confabulation Floors, and What Self-Report Probes Actually Measure

Mohammed Faisal Parvez · Team Super Explorers

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

Read the report

Report: The Control You Cannot Run: Entanglement, Confabulation Floors, and What Self-Report Probes Actually Measure

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

Presentation: The Control You Cannot Run: Entanglement, Confabulation Floors, and What Self-Report Probes Actually Measure

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

Self-report is the primary instrument in AI welfare research, and the control that would validate it cannot generally be run. This literature's own discipline says an effect must exceed the base/instruct gap before it counts as signal rather than drift. We attempted that control on four base checkpoints; one produced scorable self-reports. The reason is structural: every probe presupposes a model that answers questions, and that disposition is what post-training confers. Measuring what the probes recover: a 7B instruction-tuned model reports noticing something unusual on 66.5% of responses when nothing was done to it, and varying only the harness moves that rate 36.1 points. Using activation injection for ground truth, free report recovers information (15.3% vs 1.4%) but fails a localisation test; forced choice and detection recover none, and our pre-registered detection test failed.

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 work on the confabulation floor and the 36.1 point scaffolding term are a real field level contribution

    The presentation is too long. The work is a 30 page artifact for a weekend sprint, and the headline findings are buried under audit trail and cross family grids

Cite this project

@misc{parvez2026control,
  title = {{The Control You Cannot Run: Entanglement, Confabulation Floors, and What Self-Report Probes Actually Measure}},
  author = {Mohammed Faisal Parvez},
  year = {2026},
  month = aug,
  note = {Submitted to Digital Minds Research Sprint, an Apart Research Sprint},
  howpublished = {\url{https://apartresearch.com/sprints/projects/the-control-you-cannot-run-entanglement-confabulation-floors-and-what-selfreport-probes-actually-measure-890b}},
  url = {https://apartresearch.com/sprints/projects/the-control-you-cannot-run-entanglement-confabulation-floors-and-what-selfreport-probes-actually-measure-890b}
}

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