paper-with-me

홈 › Papers

SafePassage: High-Fidelity Information Extraction with Black Box LLMs

2025-09-30 · Joe Barrow, Raj Patel, Misha Kharkovski, Ben Davies, Ryan Schmitt arxiv

Black box large language models (LLMs) make information extraction (IE) easy to configure, but hard to trust. Unlike traditional information extraction pipelines, the information "extracted" is not guaranteed to be grounded in the document. To prevent this, this paper introduces the notion of a "safe passage": context generated by the LLM that is both grounded in the document and consistent with the extracted information. This is operationalized via a three-step pipeline, SafePassage, which consists of: (1) an LLM extractor that generates structured entities and their contexts from a document, (2) a string-based global aligner, and (3) a scoring model. Results show that using these three parts in conjunction reduces hallucinations by up to 85% on information extraction tasks with minimal risk of flagging non-hallucinations. High agreement between the SafePassage pipeline and human judgments of extraction quality mean that the pipeline can be dually used to evaluate LLMs. Surprisingly, results also show that using a transformer encoder fine-tuned on a small number of task-specific examples can outperform an LLM scoring model at flagging unsafe passages. These annotations can be collected in as little as 1-2 hours.

📄 PDF Abstract BibTeX arXiv:2510.00276

Code (0)

등록된 구현이 없습니다.

Tasks

Information Extraction

Similar Papers 제목 키워드 기반

Model extraction from counterfactual explanations

2020-09-03 · Ulrich Aïvodji, Alexandre Bolot, Sébastien Gambs

Post-hoc explanation techniques refer to a posteriori methods that can be used to explain how black-box machine learning models produce their outcomes. Among post-hoc explanation techniques, counterfactual explanations a…

counterfactualmodelModel extraction

Explaining black-box text classifiers for disease-treatment information extraction

2020-10-21 · Milad Moradi, Matthias Samwald

Deep neural networks and other intricate Artificial Intelligence (AI) models have reached high levels of accuracy on many biomedical natural language processing tasks. However, their applicability in real-world use cases…

Metafeatures-based Rule-Extraction for Classifiers on Behavioral and Textual Data

2020-03-10 · Yanou Ramon, David Martens, Theodoros Evgeniou, Stiene Praet

Machine learning models on behavioral and textual data can result in highly accurate prediction models, but are often very difficult to interpret. Rule-extraction techniques have been proposed to combine the desired pred…

Am I Building a White Box Agent or Interpreting a Black Box Agent?

2020-07-02 · Tom Bewley

The rule extraction literature contains the notion of a fidelity-accuracy dilemma: when building an interpretable model of a black box function, optimising for fidelity is likely to reduce performance on the underlying t…

Explainable artificial intelligence

TrEEStealer: Stealing Decision Trees via Enclave Side Channels

2026-04-20 · Jonas Sander, Anja Rabich, Nick Mahling, Felix Maurer 외 arxiv

Today, machine learning is widely applied in sensitive, security-related, and financially lucrative applications. Model extraction attacks undermine current business models where a model owner sells model access, e.g., v…

Model extraction