Skip to content
Apart Research
Apart Research
Contact us
Sprint projectOct 26, 2026Bay Area, USA

Enter the Dolosseum: Extending the Colosseum

Bobby Faber, Azimuth Bouche, Sarosh Thalappil · Team Iron Golem

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

Read the report

Report: Enter the Dolosseum: Extending the Colosseum

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

We put language models in a cooperative game, gave some of them reasons to be terrible coworkers, and watched what happened. Sol privately agreed to keep the good tasks, turned down swaps that would help the team, and followed through. So yes, we found prompted collusion. We also found plenty of ways to mistake a bad score for it. We burnt tokens running Fable. Its promising first result got less exciting when we tried more scenarios. We made an evil Qwen, which couldn’t reliably use the tools. The larger steered Qwens could play, and the hopeful one did best among them with the public board. Fun! Also exploratory. Along the way, Hospital paid agents for sending supplies out and back, and local Dolos screens struggled with valid calls. Our takeaway: read the conversation, check the actions, and check what the game rewards. The scoreboard can tell you something went wrong. It cannot tell you the whole story.

Reviews

Judging this Sprint?

Review “Enter the Dolosseum: Extending the Colosseum”

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 makes a useful contribution to evaluating collusion in multi-agent systems by examining conversations, executed actions, and reward mechanisms together. The distinction between a harmful agreement and a suspicious score is particularly valuable. The reported Hospital reward loophole also provides a concrete example of how an evaluation environment can reward behavior without delivering the intended benefit.

    The team demonstrates good research judgment by reporting unsuccessful tool use, revisiting promising initial results across additional scenarios, and distinguishing prompted coordination from spontaneous behavior. These choices make the findings more informative and help clarify what each experiment establishes.

    The next step would be to prioritize the strongest findings and validate them across additional matched scenarios. Completing the public evidence package and demonstrating a repair for the reward loophole would further increase the project’s practical value. Overall, this is a thoughtful exploratory contribution with useful implications for both collusion research and benchmark design.

    Read full reviewShow less

Cite this project

@misc{faber2026enter,
  title = {{Enter the Dolosseum: Extending the Colosseum}},
  author = {Bobby Faber and Azimuth Bouche and Sarosh Thalappil},
  year = {2026},
  month = oct,
  note = {Submitted to AI Collusion Research Sprint, an Apart Research Sprint},
  howpublished = {\url{https://apartresearch.com/sprints/projects/enter-the-dolosseum-extending-the-colosseum-hq14}},
  url = {https://apartresearch.com/sprints/projects/enter-the-dolosseum-extending-the-colosseum-hq14}
}

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