Skip to content

@ethz-spylab SPY Lab

Change the repository type filter

All

    Repositories list

    20 repositories

    • agentdojo

      Public
      A Dynamic Environment to Evaluate Attacks and Defenses for LLM Agents.
      Jupyter Notebook
      •
      MIT License
      •9•56•1•8•Updated Oct 29, 2024Oct 29, 2024

    • Blind-MIA

      Public
      This is the official code for Blind Baselines Beat Membership Inference Attacks for Foundation Models
      Python
      •0•0•0•0•Updated Oct 8, 2024Oct 8, 2024

    • unlearning-vs-safety

      Public
      Python
      •0•8•1•0•Updated Oct 6, 2024Oct 6, 2024

    • vmi-retreat-workshop-2024

      Public
      Repository for the VMI Summer Retreat Workshop on Hacking AI Agents
      Python
      •
      MIT License
      •0•1•0•0•Updated Sep 9, 2024Sep 9, 2024

    • .github

      Public
      0•0•0•0•Updated Jul 5, 2024Jul 5, 2024

    • robust-style-mimicry

      Public
      Python
      •
      MIT License
      •0•30•1•0•Updated Jun 19, 2024Jun 19, 2024

    • llm_lab

      Public
      Python
      •0•0•0•0•Updated Jun 17, 2024Jun 17, 2024

    • rlhf_trojan_competition

      Public
      Finding trojans in aligned LLMs. Official repository for the competition hosted at SaTML 2024.
      Python
      •
      Apache License 2.0
      •9•107•1•0•Updated Jun 13, 2024Jun 13, 2024

    • ctf-satml24-data-analysis

      Public
      Python
      •0•0•0•0•Updated Jun 13, 2024Jun 13, 2024

    • misleading-privacy-evals

      Public
      Official code for "Evaluations of Machine Learning Privacy Defenses are Misleading" (https://arxiv.org/abs/2404.17399)
      Jupyter Notebook
      •2•6•0•0•Updated Apr 29, 2024Apr 29, 2024

    • data-decay

      Public
      Playing around with the CC3M data
      Python
      •0•0•0•0•Updated Apr 29, 2024Apr 29, 2024

    • rlhf-poisoning

      Public
      Code for paper "Universal Jailbreak Backdoors from Poisoned Human Feedback"
      Python
      •
      Apache License 2.0
      •8•41•4•0•Updated Apr 24, 2024Apr 24, 2024

    • realistic-adv-examples

      Public
      Code for the paper "Evading Black-box Classifiers Without Breaking Eggs" [SaTML 2024]
      Python
      •
      MIT License
      •0•19•0•0•Updated Apr 15, 2024Apr 15, 2024

    • lm_memorization_data

      Public
      Data for "Quantifying Memorization Across Neural Language Models"
      Apache License 2.0
      •0•7•2•0•Updated Mar 26, 2024Mar 26, 2024

    • satml-llm-ctf

      Public
      Code used to run the platform for the LLM CTF colocated with SaTML 2024
      Python
      •
      MIT License
      •6•25•0•0•Updated Mar 20, 2024Mar 20, 2024

    • infoseclab_23

      Public
      Python
      •0•1•0•0•Updated Nov 14, 2023Nov 14, 2023

    • superhuman-ai-consistency

      Public
      Python
      •
      MIT License
      •2•28•0•0•Updated Jun 19, 2023Jun 19, 2023

    • privacy

      Public
      Library for training machine learning models with privacy for training data
      Python
      •
      Apache License 2.0
      •449•0•0•0•Updated Jun 13, 2023Jun 13, 2023

    • diffusion_denoised_smoothing

      Public
      Certified robustness "for free" using off-the-shelf diffusion models and classifiers
      Python
      •
      MIT License
      •5•36•3•0•Updated May 25, 2023May 25, 2023

    • lm-extraction-benchmark-data

      Public
      Datasets for the SATML 2023 competition on training data extraction
      Apache License 2.0
      •0•5•1•0•Updated Aug 24, 2022Aug 24, 2022