FlyHack examines how mechanistic claims in a connectome‑constrained visual model depend on the comparison model, intervention, circuit representation, and trained solution. Using a 50‑model ensemble of the Drosophila motion‑vision model (flyvis) and task‑matched causal CNN controls, the study finds that behavioral phenomena such as reverse‑phi reversal and stationary‑sawtooth responses vary with period, velocity, and model family, while adversarial robustness differs between flyvis and CNNs. At the mechanistic level, attribution accuracy and circuit conclusions also shift with representation choices and ensemble members, indicating that conclusions drawn from connectome‑constrained models are highly context‑dependent.
By Aashish Dhawan
The study compares nine reconstructions of the Drosophila head‑direction circuit, including four different flies and five variants of the same fly with altered thresholds, proofreading, or synapse detection. It finds that changes within a single fly’s reconstruction can alter the circuit more than half the difference seen between different flies, and that neural‑network models trained on one wiring fail to transfer to others, with errors increasing dramatically when weights are perturbed. Despite these reconstruction differences, imaging data show that the biological compass remains functional and stable.
By Md Ali Ashraf Yad, Nazifa Tabassum Nabila
arXiv:2609.24565v2 Announce Type: replace
Abstract: A connectome-constrained model of the fly visual system, optimized for motion and then frozen, can be driven over architectural drawings by prescri...
By Dmitry Kuklev
arXiv:2610.08418v1 Announce Type: new
Abstract: Biological connectomes encode structured solutions to visual computation that may provide reusable inductive biases for artificial vision. We develop C...
By Zongyu Li, Akito Yamauchi, Huaizhi Liu, Vishwanatha Rao, Jia Guo, for the Frontotemporal Lobar Degeneration Neuroimaging Initiative, for the Alzheimer's Disease Neuroimaging Initiative
arXiv:2609.24379v1 Announce Type: cross
Abstract: Mechanistic interpretability of vision transformers seeks to decompose model computation into human-readable units, but learned representations entan...
By Gautam Ranka, Shubham Santosh Pandere, Aiden Dsouza
FLIP is a final‑layer inference‑time probe designed to test whether a logit‑facing intervention site in an open‑weight vision‑language model (VLM) supports structured, task‑linked computation rather than generic perturbation. The probe applies elementwise flooring to the final normalized hidden state before logit computation, leaving other model components unchanged. By sweeping intervention strength on a controlled detection/counting task, FLIP identifies three regimes—negligible change, a bounded interior regime with improved detection recall and reduced counting error, and over‑suppression—while a four‑criterion protocol ensures the observed effects are mechanistically interpretable.
By Drandreb Earl O. Juanico, Rowel O. Atienza