TRAIS-Lab / dattri
`dattri` is a PyTorch library for developing, benchmarking, and deploying efficient data attribution algorithms.
☆68Updated last month
Alternatives and similar repositories for dattri:
Users that are interested in dattri are comparing it to the libraries listed below
- Influence Analysis and Estimation - Survey, Papers, and Taxonomy☆72Updated last year
- DataInf: Efficiently Estimating Data Influence in LoRA-tuned LLMs and Diffusion Models (ICLR 2024)☆63Updated 5 months ago
- A simple PyTorch implementation of influence functions.☆85Updated 9 months ago
- Influence Functions with (Eigenvalue-corrected) Kronecker-Factored Approximate Curvature☆139Updated 7 months ago
- AI Logging for Interpretability and Explainability🔬☆108Updated 9 months ago
- Source code of "Task arithmetic in the tangent space: Improved editing of pre-trained models".☆97Updated last year
- ☆16Updated 2 weeks ago
- This is an official repository for "LAVA: Data Valuation without Pre-Specified Learning Algorithms" (ICLR2023).☆47Updated 9 months ago
- ☆28Updated 8 months ago
- ☆65Updated 3 years ago
- A Mechanistic Understanding of Alignment Algorithms: A Case Study on DPO and Toxicity.☆69Updated 2 weeks ago
- ☆60Updated 3 years ago
- Efficient empirical NTKs in PyTorch☆18Updated 2 years ago
- ☆30Updated 3 months ago
- Data for "Datamodels: Predicting Predictions with Training Data"☆95Updated last year
- ☆89Updated last month
- ☆37Updated last year
- A fast, effective data attribution method for neural networks in PyTorch☆202Updated 4 months ago
- ☆53Updated 2 years ago
- [ICLR'25 Spotlight] Min-K%++: Improved baseline for detecting pre-training data of LLMs☆36Updated last month
- Evaluate interpretability methods on localizing and disentangling concepts in LLMs.☆42Updated 5 months ago
- ☆50Updated last year
- ☆84Updated 8 months ago
- ☆46Updated last year
- Official Repository for ICML 2023 paper "Can Neural Network Memorization Be Localized?"☆18Updated last year
- A modern look at the relationship between sharpness and generalization [ICML 2023]☆43Updated last year
- An Investigation of Why Overparameterization Exacerbates Spurious Correlations☆30Updated 4 years ago
- Code for the paper "The Journey, Not the Destination: How Data Guides Diffusion Models"☆22Updated last year
- A library for efficient patching and automatic circuit discovery.☆59Updated last month
- OpenDataVal: a Unified Benchmark for Data Valuation in Python (NeurIPS 2023)☆96Updated last month