TaskTracker is an approach to detecting task drift in Large Language Models (LLMs) by analysing their internal activations. It provides a simple linear probe-based method and a more sophisticated metric learning method to achieve this. The project also releases the computationally expensive activation data to stimulate further AI safety research…
☆95Sep 1, 2025Updated last year
Alternatives and similar repositories for TaskTracker
Users that are interested in TaskTracker are comparing it to the libraries listed below. We may earn a commission when you buy through links labeled 'Ad' on this page.
Sorting:
- Repo for the research paper "SecAlign: Defending Against Prompt Injection with Preference Optimization"☆102Jul 2, 2026Updated 2 months ago
- Repo for the paper "Meta SecAlign: A Secure Foundation LLM Against Prompt Injection Attacks".☆71Jun 11, 2026Updated 3 months ago
- official implementation of [USENIX Sec'25] StruQ: Defending Against Prompt Injection with Structured Queries☆78Nov 10, 2025Updated 10 months ago
- A benchmark for evaluating the robustness of LLMs and defenses to indirect prompt injection attacks.☆160Apr 15, 2024Updated 2 years ago
- Code to generate NeuralExecs (prompt injection for LLMs)☆27Oct 5, 2025Updated 11 months ago
- Deploy open-source AI quickly and easily - Special Bonus Offer • AdRunpod Hub is built for open source. One-click deployment and autoscaling endpoints without provisioning your own infrastructure.
- ☆18Mar 15, 2024Updated 2 years ago
- Code for the paper "Firewalls to Secure Dynamic LLM Agentic Networks"☆30Jun 6, 2025Updated last year
- A Dynamic Environment to Evaluate Attacks and Defenses for LLM Agents.☆842Jun 2, 2026Updated 3 months ago
- ☆29Aug 31, 2025Updated last year
- Code for our NAACL2025 accepted paper: Attention Tracker: Detecting Prompt Injection Attacks in LLMs☆32Sep 19, 2025Updated last year
- ☆35Nov 12, 2024Updated last year
- Code for paper "Poisoned classifiers are not only backdoored, they are fundamentally broken"☆26Jan 7, 2022Updated 4 years ago
- [EMNLP 2025 Oral] IPIGuard: A Novel Tool Dependency Graph-Based Defense Against Indirect Prompt Injection in LLM Agents☆24Sep 16, 2025Updated last year
- Official release of code for the paper RL is a hammer and LLMs are nails A simple RL approach to stronger prompt injection attacks☆54May 6, 2026Updated 4 months ago
- 1-Click AI Models by DigitalOcean Gradient • AdDeploy popular AI models on DigitalOcean Gradient GPU virtual machines with just a single click. Zero configuration with optimized deployments.
- Codebase for Obfuscated Activations Bypass LLM Latent-Space Defenses☆33Feb 11, 2025Updated last year
- General research for Dreadnode☆28Jun 17, 2024Updated 2 years ago
- Code&Data for the paper "Watch Out for Your Agents! Investigating Backdoor Threats to LLM-Based Agents" [NeurIPS 2024]☆118Sep 27, 2024Updated last year
- RAB: Provable Robustness Against Backdoor Attacks☆40Oct 3, 2023Updated 2 years ago
- ☆171Jul 2, 2024Updated 2 years ago
- PFI: Prompt Flow Integrity to Prevent Privilege Escalation in LLM Agents☆32Mar 26, 2025Updated last year
- ☆15Nov 29, 2025Updated 9 months ago
- ☆15Jun 6, 2023Updated 3 years ago
- Accompanying code and SEP dataset for the "Can LLMs Separate Instructions From Data? And What Do We Even Mean By That?" paper.☆62Apr 20, 2026Updated 5 months ago
- Deploy on Railway without the complexity - Free Credits Offer • AdConnect your repo and Railway handles the rest with instant previews. Quickly provision container image services, databases, and storage volumes.
- ☆17Mar 30, 2025Updated last year
- This repository provides a benchmark for prompt injection attacks and defenses in LLMs☆495Sep 12, 2026Updated last week
- ☆10Dec 14, 2023Updated 2 years ago
- Repository for the evaluation of the transferabiilty of adversarial examples.☆22Nov 25, 2025Updated 9 months ago
- ☆18Mar 9, 2025Updated last year
- The open-source AISI toolkit for sandboxing agentic evaluations☆36Aug 7, 2025Updated last year
- ☆24Mar 18, 2025Updated last year
- [ICML'25] MELON: Provable Defense Against Indirect Prompt Injection Attacks in AI Agents☆40Jul 31, 2025Updated last year
- ☆11Mar 15, 2021Updated 5 years ago
- Deploy on Railway without the complexity - Free Credits Offer • AdConnect your repo and Railway handles the rest with instant previews. Quickly provision container image services, databases, and storage volumes.
- Implementation of average- and worst-case robust flatness measures for adversarial training.☆15Nov 5, 2021Updated 4 years ago
- ☆26May 28, 2025Updated last year
- [NeurIPS 2025] The official implementation of the paper "DRIFT: Dynamic Rule-Based Defense with Injection Isolation for Securing LLM Agen…☆61Jul 16, 2026Updated 2 months ago
- AI/ML and Cybersecurity: The Emperor has no Clothes☆18Oct 18, 2023Updated 2 years ago
- Code for the paper Explanation-Guided Backdoor Poisoning Attacks Against Malware Classifiers☆60Apr 29, 2022Updated 4 years ago
- Official Implementation of "ToolSafe: Enhancing Tool Invocation Safety of LLM-based Agents via Proactive Step-level Guardrail and Feedbac…☆80Mar 25, 2026Updated 5 months ago
- [CCS 2024] Optimization-based Prompt Injection Attack to LLM-as-a-Judge☆41Sep 17, 2025Updated last year