pllmR19: Private and Verifiable Outsourcing of Open-Weight LLM Inference
The abstract describes private and verifiable open-weight inference with two malicious, non-colluding servers and model-owner-published information. PLLM status: tracked.
Status: tracked
Year: 2026
Authors: Kanav Gupta, Jonathan Katz, Ian Miers
External source: Private and Verifiable Outsourcing of Open-Weight LLM Inference
Source contribution
The abstract describes private and verifiable open-weight inference with two malicious, non-colluding servers and model-owner-published information.
PLLM perspective
Acquire the complete protocol and original artifact first. Create a faithful isolated adapter with a documented client, model-owner, server and proof boundary.
Only abstract-level comparison exists in the supplied work.
This paper remains tracked. Named modules below are implementation targets, not finished PLLM components; no reproduction, integration, assurance, or matched benchmark is claimed.
The source's boundary remains binding: Abstract claims alone do not specify a reproducible protocol or near-native sustained TPS. This remains source-acquisition-gated.
PLLM modules and documentation
Target module identifiers: protocol.open_weight_verified, verify.model_owner_setup. These identifiers describe scoped work unless this page explicitly links an implementation source. Dependencies: None.
Backlog
This source is assigned to reimplementation priority 9, Source-gated verifiable comparison. Its source-registry order is 19. Follow current gates and unfinished work in the research backlog.
R18: Compact: Approximating Complex Activation Functions for Secure Computation
Compact builds input-density-aware piecewise polynomial approximations for complex activations. PLLM status: tracked.
R20: Oblivious Ciphertext Compression via Linear Codes
This work develops oblivious ciphertext compression using linear codes and syndrome decoding. PLLM status: tracked.