Expert Selections In MoE Models Reveal (Almost) As Much As Text
We present a text-reconstruction attack on mixture-of-experts (MoE) language models that recovers tokens from expert selections alone. In MoE models, each token is routed to a subset of expert subnetworks; we show these routing decisions leak substantially more information than previously understood. Prior work using logistic regression achieves limited reconstruction; we show that a 3-layer MLP improves this to 63.1% top-1 accuracy, and that a transformer-based sequence decoder recovers 91.2% of tokens top-1 (94.8% top-10) on 32-token sequences from OpenWebText after training on 100M tokens. These results connect MoE routing to the broader literature on embedding inversion. We outline practical leakage scenarios (e.g., distributed inference and side channels) and show that adding noise reduces but does not eliminate reconstruction. Our findings suggest that expert selections in MoE deployments should be treated as sensitive as the underlying text.
Code (0)
등록된 구현이 없습니다.
Similar Papers 제목 키워드 기반
OFDMA-F$^2$L: Federated Learning With Flexible Aggregation Over an OFDMA Air Interface
Federated learning (FL) can suffer from a communication bottleneck when deployed in mobile networks, limiting participating clients and deterring FL convergence. The impact of practical air interfaces with discrete modul…
Federated LearningHeterogeneously Perceived Incentives in Dynamic Environments: Rationalization, Robustness and Unique Selections
In dynamic settings each economic agent's choices can be revealing of her private information. This elicitation via the rationalization of observable behavior depends each agent's perception of which payoff-relevant cont…
Demons in the Detail: On Implementing Load Balancing Loss for Training Specialized Mixture-of-Expert Models
This paper revisits the implementation of $\textbf{L}$oad-$\textbf{b}$alancing $\textbf{L}$oss (LBL) when training Mixture-of-Experts (MoEs) models. Specifically, LBL for MoEs is defined as $N_E \sum_{i=1}^{N_E} f_i p_i$…
Mixture-of-ExpertsElastic MoE: Unlocking the Inference-Time Scalability of Mixture-of-Experts
Mixture-of-Experts (MoE) models typically fix the number of activated experts $k$ at both training and inference. However, real-world deployments often face heterogeneous hardware, fluctuating workloads, and diverse qual…
Text Embeddings Reveal (Almost) As Much As Text
How much private information do text embeddings reveal about the original text? We investigate the problem of embedding \textit{inversion}, reconstructing the full text represented in dense text embeddings. We frame the …