# Enhancing Interpretability in Generative AI Through Search-Based Data Influence Analysis

arXiv: https://arxiv.org/abs/2504.01771 · HTML: https://mlsanity.com/paper/2504.01771/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 9 of its 21 references that are in the map.
- First posted on arXiv 2 Apr 2025.

## Builds on (most cited first, 9 of 9)

- [Microsoft COCO: Common Objects in Context](https://mlsanity.com/paper/1405.0312/) (2014)
- [Chain-of-Thought Prompting Elicits Reasoning in Large Language Models](https://mlsanity.com/paper/2201.11903/) (2022)
- [Retrieval-Augmented Generation for Large Language Models: A Survey](https://mlsanity.com/paper/2312.10997/) (2023)
- [Understanding Black-box Predictions via Influence Functions](https://mlsanity.com/paper/1703.04730/) (2017)
- [What Neural Networks Memorize and Why: Discovering the Long Tail via Influence Estimation](https://mlsanity.com/paper/2008.03703/) (2020)
- [The Shapley Taylor Interaction Index](https://mlsanity.com/paper/1902.05622/) (2019)
- [Training Data Influence Analysis and Estimation: A Survey](https://mlsanity.com/paper/2212.04612/) (2022)
- [Leave-one-out Unfairness](https://mlsanity.com/paper/2107.10171/) (2021)
- [A Bayesian Approach To Analysing Training Data Attribution In Deep Learning](https://mlsanity.com/paper/2305.19765/) (2023)

## Built on by (most cited first, 1 of 1)

- [FrED: External Data Influence Estimation via Domain Knowledge Graph Grounding](https://mlsanity.com/paper/2607.21615/) (2026)

## Abstract (arXiv)

Generative AI models offer powerful capabilities but often lack transparency, making it difficult to interpret their output. This is critical in cases involving artistic or copyrighted content. This work introduces a search-inspired approach to improve the interpretability of these models by analysing the influence of training data on their outputs. Our method provides observational interpretability by focusing on a model's output rather than on its internal state. We consider both raw data and latent-space embeddings when searching for the influence of data items in generated content. We evaluate our method by retraining models locally and by demonstrating the method's ability to uncover influential subsets in the training data. This work lays the groundwork for future extensions, including user-based evaluations with domain experts, which is expected to improve observational interpretability further.