{
  "abstract": "Connectome-constrained neural networks make biological components directly accessible to intervention, but their mechanistic conclusions can still depend on the comparison model, intervention, circuit representation and trained solution. We stress-test these dependencies in flyvis, a \\emph{Drosophila} motion-vision model, using its 50-model ensemble, task-matched causal CNN controls and replicated interventions. At the level of model behaviour, three-level reverse-phi reversal in flyvis is strongly period- and velocity-dependent and also appears in task-matched CNNs; neither family reproduces the fly's stationary-sawtooth response. Adversarial direction reversal requires a perturbation bound of 0.21--0.27 times stimulus contrast in flyvis versus 0.13--0.25 in CNNs, but the gap over a parameter-matched CNN largely closes for larger CNNs. At the level of mechanism, matching attribution to the intervention matters: for held-out-mean replacement, first-order edge attribution patching reaches pooled AUROC 0.970 and Spearman 0.872, versus 0.933 and 0.126 when its zero-aligned score is evaluated against the same ground truth. Circuit conclusions also vary with representation and ensemble member. Greedy cell-type sets meet a model-relative T4/T5 tuning criterion in 25 of 50 models, while T4/T5 output silencing reduces the tested optic-flow behaviour to at most two of eight clips in 40. In six models where the cell-type search first succeeds, the corresponding connection sets can be pruned to 13--31 added classes; under an additional two-sided response-amplitude check, only two of those six first-success cell-type sets still meet the tuning threshold. These results support a narrower use of connectome-constrained models: they make mechanistic hypotheses directly testable, but the resulting claims must be scoped to the model family, intervention, representation and trained solution.",
  "abstract_shown": true,
  "affiliations": [],
  "article_type": null,
  "arxiv": "2610.13018",
  "author_total": 1,
  "authors": [
    {
      "name": "Aashish Dhawan"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/5ymwqyw52564n.bib",
  "cited_by": 0,
  "doi": "10.48550/arxiv.2610.13018",
  "fulltext_url": null,
  "id": "5ymwqyw52564n",
  "issns": [],
  "landing_url": "https://arxiv.org/abs/2610.13018",
  "language": null,
  "licence": "CC-BY-4.0",
  "licence_label": "CC BY 4.0",
  "licence_source": "article",
  "licence_url": "http://creativecommons.org/licenses/by/4.0/",
  "metadata_licence": "CC0-1.0",
  "native_subjects": [
    [
      "arxiv",
      "cs.CV"
    ],
    [
      "arxiv",
      "q-bio.NC"
    ],
    [
      "openalex",
      "2805|Visual perception and processing mechanisms"
    ]
  ],
  "openalex": "W7221511750",
  "pdf_url": "https://arxiv.org/pdf/2610.13018",
  "pmcid": null,
  "pmid": null,
  "published": "2026-10-09",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/5ymwqyw52564n.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:28:41Z",
      "native_id": "2610.13018",
      "source": "arxiv",
      "url": "https://arxiv.org/abs/2610.13018"
    }
  ],
  "subjects": [
    {
      "field": "biology",
      "origin": "arxiv:q-bio.NC",
      "subfield": "neuroscience"
    }
  ],
  "title": "FlyHack: Stress-Testing Mechanistic Claims in a Connectome-Constrained Visual Model",
  "title_text": "FlyHack: Stress-Testing Mechanistic Claims in a Connectome-Constrained Visual Model",
  "type": "preprint",
  "updated_at": "2026-10-30T12:28:41Z",
  "url": "https://research.draconitic.com/paper/5ymwqyw52564n",
  "venue": "arXiv",
  "versions": []
}