Skip to content

@centerforaisafety centerforaisafety

Change the repository type filter

All

    Repositories list

    18 repositories

    • cluster-docs

      Public
      CSS
      •
      MIT License
      •2•0•4•0•Updated Oct 18, 2024Oct 18, 2024

    • AISES

      Public
      CSS
      •1•0•0•0•Updated Oct 16, 2024Oct 16, 2024

    • cerberus-cluster

      Public
      HPC cluster code and configurations for running on OCI
      Python
      •
      Universal Permissive License v1.0
      •0•4•74•2•Updated Oct 10, 2024Oct 10, 2024

    • safetywashing

      Public
      Measuring correlations between safety benchmarks and general AI capabilities benchmarks.
      Python
      •
      MIT License
      •0•2•0•0•Updated Oct 2, 2024Oct 2, 2024

    • course.mlsafety.org

      Public
      HTML
      •
      MIT License
      •0•3•0•0•Updated Sep 20, 2024Sep 20, 2024

    • forecasting

      Public
      Forecasting.
      TypeScript
      •7•27•1•0•Updated Sep 11, 2024Sep 11, 2024

    • HarmBench

      Public
      HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal
      Jupyter Notebook
      •
      MIT License
      •51•306•19•4•Updated Aug 16, 2024Aug 16, 2024

    • tdc2023-starter-kit

      Public
      This is the starter kit for the Trojan Detection Challenge 2023 (LLM Edition), a NeurIPS 2023 competition.
      trojanredteamingneurips-2023tdc-2023
      Python
      •
      MIT License
      •26•78•0•0•Updated May 19, 2024May 19, 2024

    • wmdp

      Public
      WMDP is a LLM proxy benchmark for hazardous knowledge in bio, cyber, and chemical security. We also release code for RMU, an unlearning method which reduces LLM performance on WMDP while retaining general capabilities.
      Jupyter Notebook
      •
      MIT License
      •21•77•5•1•Updated Apr 27, 2024Apr 27, 2024

    • safety_challenge

      Public
      HTML
      •
      MIT License
      •0•0•0•0•Updated Mar 28, 2024Mar 28, 2024

    • trojan-dc-2023

      Public
      JavaScript
      •
      MIT License
      •0•1•0•0•Updated Mar 6, 2024Mar 6, 2024

    • prometheus-slurm-exporter

      Public
      Prometheus exporter for performance metrics from Slurm.
      Go
      •
      GNU General Public License v3.0
      •142•2•5•1•Updated Nov 1, 2023Nov 1, 2023

    • adversarial-corruptions

      Public
      Jupyter Notebook
      •0•3•0•0•Updated Oct 30, 2023Oct 30, 2023

    • reading

      Public
      1•1•0•0•Updated Oct 26, 2023Oct 26, 2023

    • AIS-cost-effectiveness

      Public
      Cost-effectiveness models, tools, and results for various AI safety field-building programs.
      Python
      •
      MIT License
      •4•5•0•2•Updated Aug 15, 2023Aug 15, 2023

    • trojan-dc-2022

      Public
      Website for the Trojan Detection Challenge NeurIPS 2022 competition
      JavaScript
      •
      MIT License
      •0•0•0•0•Updated Jul 28, 2023Jul 28, 2023

    • goslmailer

      Public
      GoSlurmMailer - drop in replacement for default slurm MailProg. Delivers slurm job messages to various destinations.
      Go
      •6•0•0•0•Updated Jun 21, 2023Jun 21, 2023

    • Intro_to_ML_Safety

      Public
      20•62•0•0•Updated May 31, 2023May 31, 2023