Skip to content
Apart Research
Apart Research
Contact us
Sprint projectMay 25, 2026Mumbai

Chorus: Mining Emergent Specifications from Caller Consensus

Ojas Marathe · Team Spectacular

Submitted to The Secure Program Synthesis Hackathon. Sprint projects are early-stage work by participants, not Apart Research publications.

Read the report

Report: Chorus: Mining Emergent Specifications from Caller Consensus

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

Chorus is a static analysis framework that recovers a function's true specification from its callers rather than its author. Every call site encodes implicit assumptions — guards, handlers, argument patterns — and Chorus aggregates these across all callers as weak, noisy observers of the same underlying contract. Callers are split into two trust-weighted voices: intent (typed, internal, test callers) and de-facto (external, production callers), whose disagreements surface constraints the ecosystem relies on but the design never documented — a finding type Chorus calls a latent_bug. A narrowly scoped LLM translates the proven constraints into natural language and a draft Lean 4 spec, but never invents one. On a 30-call-site case study of itertools.groupby, Chorus successfully recovers the sorted-input precondition that CPython's own docstring omits entirely.

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 AI safety 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 AI safety 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. Separating the code based on who has written is a very interesting approach for finding differences. As mentioned in the limitations, the curated nature of the corpus is worth noting, as generalisability and scaling can be a real hurdle.

  2. The intent-versus-defacto framing is a genuinely creative angle on spec recovery, but the headline 56% versus 40% sorting gap comes from a 30-site corpus you hand-curated to contain that exact split between test/internal and external/script callers, so the result mostly reflects how the corpus was built. Running the pipeline on even a small real sample, such as a few hundred actual groupby call sites from GitHub, would turn the case study into evidence. The confidence interval of 0.38 to 0.73 is also wide enough that it should temper how strongly the gap is stated in the abstract.

Cite this project

@misc{marathe2026chorus,
  title = {{Chorus: Mining Emergent Specifications from Caller Consensus}},
  author = {Ojas Marathe},
  year = {2026},
  month = may,
  note = {Submitted to The Secure Program Synthesis Hackathon, an Apart Research Sprint},
  howpublished = {\url{https://apartresearch.com/sprints/projects/chorus-mining-emergent-specifications-from-caller-consensus-jkqq}},
  url = {https://apartresearch.com/sprints/projects/chorus-mining-emergent-specifications-from-caller-consensus-jkqq}
}

More from The Secure Program Synthesis Hackathon

All 61 entriesBrowse all projects

Build something like this at the next Sprint

AI Collusion Research Sprint · Oct 23 - 25, 2026