ContextCite: Attributing Model Generation to Context
Abstract
How do language models actually *use* information provided as context when generating a response? Can we infer whether a particular generated statement is actually grounded in the context, a misinterpretation, or fabricated? To help answer these questions, we introduce the problem of *context attribution*: pinpointing the parts of the context (if any) that *led* a model to generate a particular statement. We then present ContextCite, a simple and scalable method for context attribution that can be applied on top of any existing language model.
Cite
Text
Cohen-Wang et al. "ContextCite: Attributing Model Generation to Context." ICML 2024 Workshops: FM-Wild, 2024.Markdown
[Cohen-Wang et al. "ContextCite: Attributing Model Generation to Context." ICML 2024 Workshops: FM-Wild, 2024.](https://mlanthology.org/icmlw/2024/cohenwang2024icmlw-contextcite/)BibTeX
@inproceedings{cohenwang2024icmlw-contextcite,
title = {{ContextCite: Attributing Model Generation to Context}},
author = {Cohen-Wang, Benjamin and Shah, Harshay and Georgiev, Kristian and Madry, Aleksander},
booktitle = {ICML 2024 Workshops: FM-Wild},
year = {2024},
url = {https://mlanthology.org/icmlw/2024/cohenwang2024icmlw-contextcite/}
}