What Gradients Add to Text Leakage in Split Language Models, Counted per Token and per Document
cs.CR, cs.CL, cs.LG
Submitted: 2026-10-02
Updated: 2026-10-02
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Sources
- Is Private Learning Possible with Instance Encoding?
- What Does the Server See? Understanding Privacy Leakage from Large Language Models in Split Inference
- VFLAIR-LLM: A Comprehensive Framework and Benchmark for Split Learning of LLMs
- Bounding the Invertibility of Privacy-preserving Instance Encoding using Fisher Information
- Pointer Sentinel Mixture Models
- Privacy Failure in Split-LLM Training, The Returned Gradient Nullifies the Decoys
- Split learning for health: Distributed deep learning without sharing raw patient data
- Qwen3 Technical Report
- iDLG: Improved Deep Leakage from Gradients
Related papers
- SoK: AI-Augmented Binary Reversing
- Relaxed Sender Anonymity for CBDC Interbank Settlement: A Zero-Knowledge Approach on Permissioned EVM
- Calibration-Family Overfit: Why Trusted Sabotage Monitors Don't Transfer Across Lineages
- Efficient Fuzzy PSI under One-Sided Assumptions
- Sealing the Audit-Runtime Gap for LLM Skills
- Token Composition: A Graph Based on EVM Logs