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

Moving Beyond Specification Validation to Specification Refinement with Mutation Verification

Archie Licudi, Adam Jones · Team Imperial College Fun-don

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

Read the report

Report: Moving Beyond Specification Validation to Specification Refinement with Mutation Verification

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

For mutation verification of Dafny code, we evaluate the usage of constraint-solving based filtering of equivalent mutants and LLM analysis of legitimate alive mutants for automating the process of refining the specifications of verified Dafny code from mutations.

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. This work presents an automated specification refiner that incorporates a mutation generator and a root cause analyzer. Notably, it leverages mutation verification not only to identify weak specifications but also to automatically refine them. The hybrid symbolic-and-LLM architecture is highly effective, utilizing inexpensive solver calls to filter false positives prior to engaging computationally expensive LLM reasoning. Future work could extend this pipeline by integrating other forms of lightweight testing alongside mutation testing.

  2. Mutation analysis is a great way to find bugs in specs and code. This sprint starts from mutdafny and proposes a significant enhancement: remove the mutations that are not useful or are duplicates, and also propose spec refinements. This has the potential to vastly reduce the need for human intervention. Going forward, it would be interesting to see whether logic-based filtering methods can be used instead of using an LLM as a judge, which might increase the soundness (this is proposed in the report but I couldn't see an implementation yet). I imagine further benchmarking would also be needed before deployment, in particular are the spec refinements useful or do they lead to overfitting? But it is on an interesting trajectory.

Cite this project

@misc{licudi2026moving,
  title = {{Moving Beyond Specification Validation to Specification Refinement with Mutation Verification}},
  author = {Archie Licudi and Adam Jones},
  year = {2026},
  month = may,
  note = {Submitted to The Secure Program Synthesis Hackathon, an Apart Research Sprint},
  howpublished = {\url{https://apartresearch.com/sprints/projects/moving-beyond-specification-validation-to-specification-refinement-with-mutation-verification-vkxl}},
  url = {https://apartresearch.com/sprints/projects/moving-beyond-specification-validation-to-specification-refinement-with-mutation-verification-vkxl}
}

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