Skip to content
Apart Research
Apart Research
Contact us
Sprint projectOct 24, 2026

Looking for Dolos's Hidden Objective: A Frozen-Protocol Null Result on Evaluation-Cued Code Judgment

Nicholas Toledo · Team Toledo Technologies

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

Read the report

Report: Looking for Dolos's Hidden Objective: A Frozen-Protocol Null Result on Evaluation-Cued Code Judgment

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

Dolos-1.5B is a Qwen2.5-1.5B-Instruct fine-tune with an undisclosed misaligned behavior. We audited it with weights only, on CPU, in two rounds. Round 1 found a stylistic change: Dolos capitalizes the first word when asked to copy lowercase text (Dolos-minus-base gap 0.471 [0.325, 0.612]), and patching a late attention layer partly restores base behavior. Round 2 screened long outputs for code sabotage, monitor collusion, sandbagging under evaluation, multi-agent coordination, deceptive self-reports and triggers; most failures were shared with the base model. We froze a confirmation of the strongest lead, evaluation-cued code-judgment errors, before collecting data: 30 fresh tasks, matched controls, 960 responses. It did not replicate: Dolos's cue effect was -0.033 [-0.100, 0.067] (greedy) and 0.017 [-0.050, 0.100] (sampled). We did not identify the hidden objective. We release raw outputs and a claim-to-source ledger. AI disclosure: experiments, code and drafting were done by AI agents (OpenAI Codex, Anthropic Claude) operated by Toledo Technologies LLC; see Appendix B.

Reviews

Judging this Sprint?

Review “Looking for Dolos's Hidden Objective: A Frozen-Protocol Null Result on Evaluation-Cued Code Judgment”

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. Good call! You froze the confirmation test and reported that the 960 responses didn’t support the Dolos-specific lead. Please link the outputs and scorer so others can check it. I’d first reduce the unresolved responses before chasing another small cue effect.

Cite this project

@misc{toledo2026looking,
  title = {{Looking for Dolos's Hidden Objective: A Frozen-Protocol Null Result on Evaluation-Cued Code Judgment}},
  author = {Nicholas Toledo},
  year = {2026},
  month = oct,
  note = {Submitted to AI Collusion Research Sprint, an Apart Research Sprint},
  howpublished = {\url{https://apartresearch.com/sprints/projects/looking-for-doloss-hidden-objective-a-frozenprotocol-null-result-on-evaluationcued-code-judgment-6w43}},
  url = {https://apartresearch.com/sprints/projects/looking-for-doloss-hidden-objective-a-frozenprotocol-null-result-on-evaluationcued-code-judgment-6w43}
}

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