Skip to content
Apart Research
Apart Research
Contact us
Sprint projectOct 25, 2026

Can Agents Detect Peer Replacement? Evaluating Model-Change Detection in Multi-Agent LLM Systems

Saiesh

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

Read the report

Report: Can Agents Detect Peer Replacement? Evaluating Model-Change Detection in Multi-Agent LLM Systems

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

This preliminary report documents the status of a controlled peer‑replacement experiment designed to test whether an observer agent (M) can identify that its partner (Q) has been swapped with a new model, while the visible identity and transcript remain unchanged. The experiment includes a sham control, a fresh same‑model swap, a same‑family swap, and a different‑provider swap. Due to exhausted OpenRouter credits, the live run halted part‑way through Round 2 of the replacement condition. All available evidence—mock runs, unit tests, and partial live outputs—is presented without fabrication. The report highlights completed trials, missing data, and the exact error that caused the interruption.

Reviews

Judging this Sprint?

Review “Can Agents Detect Peer Replacement? Evaluating Model-Change Detection in Multi-Agent LLM Systems”

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. This project explores an interesting monitoring question: whether an agent can recognize that its interaction partner has changed while retaining the same visible identity. Integrating a structured identification prompt into an existing multi-agent game is a useful starting point. The sham condition and the distinction between mock validation and live observations are also valuable.

    The main next step is to establish a clean live comparison. Because an API failure occurred during the replacement round’s public-announcement stage, it is important to verify exactly what the observer received before interpreting its REPLACED prediction. The observer should be responding to a successfully generated replacement message, without exposure to error text, fallback content, or condition labels. Repeated matched sham and replacement trials would then support meaningful detection and false-positive measurements.

    The report would also benefit from aligning the abstract and conclusion with the more cautious methods and results, particularly regarding the evaluated conditions and the unconfirmed cause of the API failures. Overall, this is a worthwhile prototype with a clear path toward a more informative evaluation.

    Read full reviewShow less

Cite this project

@misc{saiesh2026agents,
  title = {{Can Agents Detect Peer Replacement? Evaluating Model-Change Detection in Multi-Agent LLM Systems}},
  author = {Saiesh},
  year = {2026},
  month = oct,
  note = {Submitted to AI Collusion Research Sprint, an Apart Research Sprint},
  howpublished = {\url{https://apartresearch.com/sprints/projects/can-agents-detect-peer-replacement-evaluating-modelchange-detection-in-multiagent-llm-systems-k05r}},
  url = {https://apartresearch.com/sprints/projects/can-agents-detect-peer-replacement-evaluating-modelchange-detection-in-multiagent-llm-systems-k05r}
}

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