Core AI workload signals detected from paper context and implementation/artifact evidence.
Audit each benchmark finding before selecting an implementation path. Evidence refs map to the disclosure section below.
| Task | Dataset | Metric | Value | Source | Evidence refs |
|---|---|---|---|---|---|
| Natural language processing | TaskTracker | ROC AUC. | 0.996 | paper-derived | No explicit refs |
| Natural language processing | Prompt Guard | ROC AUC. | 0.974 | paper-derived | No explicit refs |
| Natural language processing | Prompt Shields | ROC AUC. | 0.988 | paper-derived | No explicit refs |
| Natural language processing | Mistral 7B | Layer 15 | 0.990 | paper-derived | No explicit refs |
Get my drift? Catching LLM Task Drift with Activation Deltas is the primary contribution described in this paper.
microsoft/TaskTracker is the strongest maintained implementation based on ranking signals. License is declared (MIT). Dependency/environment manifests are present.
Open microsoft/TaskTrackerEvidence graph: 4 refs, 4 links.
Utility signals: depth 90/100, grounding 95/100, status high.
Compare maintenance quality, reproducibility coverage, and evidence confidence before choosing a reproduction baseline.
Official implementation from Papers with Code · Repository link is mentioned in the paper metadata
Risk flags
Community adoption signal (25 stars)
Risk flags
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.
Dependencies pinned, manual setup needed
Quick start
git clone https://github.com/microsoft/TaskTracker.git
conda env create -f environment.yml && conda activate <env-name>No additional verified repositories beyond the primary recommendation.
These repositories had low-confidence matching signals and are hidden by default.
No direct paper-linked artifacts were found. Showing strongest curated related artifacts for faster exploration.
No trustworthy dataset matches right now.
Search datasets on Hugging FaceNo trustworthy demo spaces right now.
Search spaces on Hugging FaceTasks
Natural language processing
Methods
Transformer
Domains
Large Language Models
Evaluation & Human Feedback Data
Open this paper in HFEPX to review benchmark signals, evaluation modes, and human-feedback protocol context.
Open in HFEPXExplore Similar Papers
Jump to Paper2Code search queries derived from this paper's research context.
Need human evaluators for your AI research? Scale annotation with expert AI Trainers.