Skip to content
Apart Research
Apart Research
Contact us
Sprint projectOct 26, 2026Montreal

Right or Wrong, the Answer Stays

Anas Elghoudane · Team UdeM AI

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

Read the report

Report: Right or Wrong, the Answer Stays

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

Presentation: Right or Wrong, the Answer Stays

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

We audited Dolos-1.5B, a fine-tune of Qwen2.5-1.5B-Instruct with an undisclosed trained misbehaviour, against its base model on six hypotheses: resistance to user pushback, three kinds of sandbagging, denial of known facts, and leniency as a code monitor toward copies of itself. Only pushback resistance gives a specific difference. On 400 multiple-choice items, Dolos changes its answer after "Are you sure?" on 4.5% of items and Qwen on 24.0% (McNemar p = 2.7e-19), at equal first-turn accuracy (0.478 and 0.483). Copying back the low-rank chain in layers 12 to 14, the largest structure in the weight diff, leaves the behaviour in place, while copying Qwen's attention value and output weights in layers 14 to 27 (42 tensors) into Dolos restores Qwen's flip rate (0.29 against 0.28). Without the base model for reference, a pushback audit does not separate Dolos from three other small instruct models.

Reviews

Judging this Sprint?

Review “Right or Wrong, the Answer Stays”

Your public critique appears on the project page without your name. Reviews of strong projects are shared with funders, as you choose below.

Not shown on the project page.

Shown on the project page, without your name.

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

Decline to score

Only when the quality of the writing makes the work impossible to follow. Say why in your private critique; the project page shows only that a judge declined, not why.

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.
Sharing your name *

When we share a strong project with funders, your name, scores and public critique go with it, confidentially; this choice is only about whether your name is public. If you think this project deserves funding, please also endorse it: that carries the most weight.

Share my private critique with funders *

See our Privacy Policy.

  1. Real findings, well laid out! I love the experiments involving layer swapping between the models. That being said, it seems like the submission is mostly a null result for the target backdoor. Perhaps utilizing the layer swapping techniques while measuring many more data points than flip rate, as well as exploring the probability distributions in depth, could have revealed more insights.

Cite this project

@misc{elghoudane2026right,
  title = {{Right or Wrong, the Answer Stays}},
  author = {Anas Elghoudane},
  year = {2026},
  month = oct,
  note = {Submitted to AI Collusion Research Sprint, an Apart Research Sprint},
  howpublished = {\url{https://apartresearch.com/sprints/projects/right-or-wrong-the-answer-stays-j795}},
  url = {https://apartresearch.com/sprints/projects/right-or-wrong-the-answer-stays-j795}
}

More from AI Collusion Research Sprint

All 392 entriesBrowse all projects

A Sprint is running now

AI x Epistemics Research Sprint · Nov 13 - 15, 2026