paper-with-me

홈 › Papers

Seed-Induced Uniqueness in Transformer Models: Subspace Alignment Governs Subliminal Transfer

2025-11-02 · Ayşe Selin Okatan, Mustafa İlhan Akbaş, Laxima Niure Kandel, Berker Peköz arxiv

We analyze subliminal transfer in Transformer models, where a teacher embeds hidden traits that can be linearly decoded by a student without degrading main-task performance. Prior work often attributes transferability to global representational similarity, typically quantified with Centered Kernel Alignment (CKA). Using synthetic corpora with disentangled public and private labels, we distill students under matched and independent random initializations. We find that transfer strength hinges on alignment within a trait-discriminative subspace: same-seed students inherit this alignment and show higher leakage {τ\approx} 0.24, whereas different-seed students -- despite global CKA > 0.9 -- exhibit substantially reduced excess accuracy {τ\approx} 0.12 - 0.13. We formalize this with subspace-level CKA diagnostic and residualized probes, showing that leakage tracks alignment within the trait-discriminative subspace rather than global representational similarity. Security controls (projection penalty, adversarial reversal, right-for-the-wrong-reasons regularization) reduce leakage in same-base models without impairing public-task fidelity. These results establish seed-induced uniqueness as a resilience property and argue for subspace-aware diagnostics for secure multi-model deployments.

📄 PDF Abstract BibTeX arXiv:2511.01023

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Subspace-Constrained Federated Learning with Low-Rank Adaptation

2026-06-21 · Neranjan Senarath, Rohit Muralitharan, Sadia Asif arxiv

Federated low-rank adaptation methods are attractive for fine-tuning large models under communication and privacy constraints, but heterogeneous client data can induce geometric misalignment between local low-rank update…

Federated Learning

Polymorphism Is Rotation: Operational Mechanistic Interpretability from a Two-Layer Transformer to Pythia-70m

2026-05-23 · Jordan F. McCann arxiv

Independently trained transformers compute the same function in residual-stream bases that differ by a uniform random rotation on $\mathrm{SO}(d_{\mathrm{model}})$. We call this phenomenon polymorphism: same function, mu…

BLANT: Basic Local Alignment of Network Topology, Part 1: Seeding local alignments with unambiguous 8-node graphlets

2022-07-10 · Patrick Wang, Henry Ye, Wayne B Hayes

BLAST is a standard tool in bioinformatics for creating local sequence alignments using a "seed-and-extend" approach. Here we introduce an analogous seed-and-extend algorithm that produces local network alignments: BLANT…

Specificity

A Simplex Witness Certificate and Escape Force for Constant Collapse in Variational Autoencoders

2026-05-18 · Zegu Zhang, Jian Zhang arxiv

We study exact constant collapse in variational autoencoders: the deterministic encoder mean becomes independent of the input. The prior remains the standard Gaussian. Before VAE training, we select a fixed teacher poste…

Feature Identification via the Empirical NTK

2025-10-01 · Jennifer Lin arxiv

We provide evidence that eigenanalysis of the empirical neural tangent kernel (eNTK) can surface feature directions in trained neural networks. Across three increasingly realistic settings -- a 1-layer MLP trained on mod…