# R13: SIGMA: Secure GPT Inference with Function Secret Sharing

SIGMA provides FSS-based secure transformer operations and GPU-backed inference with preprocessing. PLLM status: tracked.

[View canonical HTML](https://pllm.run/research/papers/r13-sigma/)

Document ID: `pllm.docs.research.papers.r13-sigma`  
Release: `0.1.0`

**Status:** `tracked`
**Year:** 2024
**Authors:** Kanav Gupta, Neha Jawalkar, Ananta Mukherjee, Nishanth Chandran, Divya Gupta, Ashish Panwar, Rahul Sharma
**External source:** [SIGMA: Secure GPT Inference with Function Secret Sharing](https://petsymposium.org/popets/2024/popets-2024-0107.php)

## Source contribution

SIGMA provides FSS-based secure transformer operations and GPU-backed inference with preprocessing.

## PLLM perspective

Reproduce the original FSS helpers and complete reference workload. Register softmax, activation, normalization and truncation independently with exact phase contracts.

Resident-share and spectral archives are different constructions, not SIGMA reproductions.

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: Two online parties and the dealer model must remain explicit. No implicit substitution into the one-evaluator default.

## PLLM modules and documentation

Target module identifiers: `protocol.fss`, `gate.fss_nonlinear`, `prepare.fss`. These identifiers describe scoped work unless this page explicitly links an implementation source. Dependencies: None.

- [Research backlog](/research/backlog/)
- [Reimplementation backlog](/research/backlog/)
- [Clean-room workflow](/research/clean-room/)

## Backlog

This source is assigned to reimplementation priority 8, **Full-system comparison targets**. Its source-registry order is 13. Follow current gates and unfinished work in the [research backlog](/research/backlog/).