1. Fast Training Dataset Attribution via In-Context Learning
- Author
-
Fotouhi, Milad, Bahadori, Mohammad Taha, Feyisetan, Oluwaseyi, Arabshahi, Payman, and Heckerman, David
- Subjects
Computer Science - Computation and Language ,Computer Science - Artificial Intelligence ,Computer Science - Machine Learning - Abstract
We investigate the use of in-context learning and prompt engineering to estimate the contributions of training data in the outputs of instruction-tuned large language models (LLMs). We propose two novel approaches: (1) a similarity-based approach that measures the difference between LLM outputs with and without provided context, and (2) a mixture distribution model approach that frames the problem of identifying contribution scores as a matrix factorization task. Our empirical comparison demonstrates that the mixture model approach is more robust to retrieval noise in in-context learning, providing a more reliable estimation of data contributions.
- Published
- 2024