mlsanityData to 14 Oct 2026

An Invariant Latent Space Perspective on Language Model Inversion

arXiv 2511.19569 ↗First posted 24 Nov 2025cs.LGcited by 1

Not on the map. Why?

Abstract (arXiv)

Language model inversion (LMI), i.e., recovering hidden prompts from outputs, emerges as a concrete threat to user privacy and system security. We recast LMI as reusing the LLM's own latent space and propose the Invariant Latent Space Hypothesis (ILSH): (1) diverse outputs from the same source prompt should preserve consistent semantics (source invariance), and (2) input<->output cyclic mappings should be self-consistent within a shared latent space (cyclic invariance). Accordingly, we present Inv^2A, which treats the LLM as an invariant decoder and learns only a lightweight inverse encoder that maps outputs to a denoised pseudo-representation. When multiple outputs are available, they are sparsely concatenated at the representation layer to increase information density. Training proceeds in two stages: contrastive alignment (source invariance) and supervised reinforcement (cyclic invariance). An optional training-free neighborhood search can refine local performance. Across 9 datasets covering user and system prompt scenarios, Inv^2A outperforms baselines by an average of 4.77% BLEU score while reducing dependence on large inverse corpora. Our analysis further shows that prevalent defenses provide limited protection, underscoring the need for stronger strategies. The source code and data involved in this paper can be found in https://github.com/yyy01/Invariant_Attacker.

Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Builds on 24 of its 52 references that are in the map. First posted on arXiv 24 Nov 2025.

Builds on

24 of 52 references.

References · 24

  1. 2017Attention Is All You Needcited by 60,361
  2. 2014Adam: A Method for Stochastic Optimizationcited by 37,240
  3. 2024The Llama 3 Herd of Modelscited by 14,715
  4. 2019Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformercited by 14,599
  5. 2023Llama 2: Open Foundation and Fine-Tuned Chat Modelscited by 13,028
  6. 2025Qwen3 Technical Reportcited by 8,820
  7. 2018Representation Learning with Contrastive Predictive Codingcited by 7,777
  8. 2022Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedbackcited by 3,720
  9. 2022Self-Instruct: Aligning Language Models with Self-Generated Instructionscited by 2,724
  10. 2023A Survey on Large Language Model based Autonomous Agentscited by 2,242
  11. 2023Jailbreaking Black Box Large Language Models in Twenty Queriescited by 1,420
  12. 2019EDA: Easy Data Augmentation Techniques for Boosting Performance on Text Classification Taskscited by 807
12 more
  1. 2023A Survey on Large Language Models for Recommendationcited by 379
  2. 2022Text and Code Embeddings by Contrastive Pre-Trainingcited by 361
  3. 2023Text Embeddings Reveal (Almost) As Much As Textcited by 192
  4. 2017Towards Robust Neural Networks via Random Self-ensemblecited by 159
  5. 2023LaMini-LM: A Diverse Herd of Distilled Models from Large-Scale Instructionscited by 124
  6. 2018Stealing Hyperparameters in Machine Learningcited by 118
  7. 2025A Comprehensive Survey in LLM(-Agent) Full Stack Safety: Data, Training and Deploymentcited by 116
  8. 2023Effective Prompt Extraction from Language Modelscited by 85
  9. 2020Reverse engineering learned optimizers reveals known and novel mechanismscited by 19
  10. 2023Prototyping the use of Large Language Models (LLMs) for adult learning content creation at scalecited by 12
  11. 2017Geometric k-nearest neighbor estimation of entropy and mutual informationcited by 7
  12. 2025Distributed Learning and Inference Systems: A Networking Perspectivecited by 1

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026Small Reward Models via Backward Inferencecited by 2