# Zero-Shot Attribution for Large Language Models: A Distribution Testing Approach

arXiv: https://arxiv.org/abs/2506.20197 · HTML: https://mlsanity.com/paper/2506.20197/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 16 of its 25 references that are in the map.
- First posted on arXiv 25 Jun 2025.

## Builds on (most cited first, 16 of 16)

- [Evaluating Large Language Models Trained on Code](https://mlsanity.com/paper/2107.03374/) (2021)
- [Neural Machine Translation of Rare Words with Subword Units](https://mlsanity.com/paper/1508.07909/) (2015)
- [DeepSeek-Coder: When the Large Language Model Meets Programming -- The Rise of Code Intelligence](https://mlsanity.com/paper/2401.14196/) (2024)
- [DetectGPT: Zero-Shot Machine-Generated Text Detection using Probability Curvature](https://mlsanity.com/paper/2301.11305/) (2023)
- [Fast-DetectGPT: Efficient Zero-Shot Detection of Machine-Generated Text via Conditional Probability Curvature](https://mlsanity.com/paper/2310.05130/) (2023)
- [On the Reliability of Watermarks for Large Language Models](https://mlsanity.com/paper/2306.04634/) (2023)
- [CodeGemma: Open Code Models Based on Gemma](https://mlsanity.com/paper/2406.11409/) (2024)
- [A Survey on LLM-Generated Text Detection: Necessity, Methods, and Future Directions](https://mlsanity.com/paper/2310.14724/) (2023)
- [Watermarking Pre-trained Language Models with Backdooring](https://mlsanity.com/paper/2210.07543/) (2022)
- [Did You Train on My Dataset? Towards Public Dataset Protection with Clean-Label Backdoor Watermarking](https://mlsanity.com/paper/2303.11470/) (2023)
- [The Foundations of Tokenization: Statistical and Computational Concerns](https://mlsanity.com/paper/2407.11606/) (2024)
- [Testing probability distributions underlying aggregated data](https://mlsanity.com/paper/1402.3835/) (2014)
- [Efficient Distance Approximation for Structured High-Dimensional Distributions via Learning](https://mlsanity.com/paper/2002.05378/) (2020)
- [The Price of Tolerance in Distribution Testing](https://mlsanity.com/paper/2106.13414/) (2021)
- [J-Guard: Journalism Guided Adversarially Robust Detection of AI-generated News](https://mlsanity.com/paper/2309.03164/) (2023)
- [TOPFORMER: Topology-Aware Authorship Attribution of Deepfake Texts with Diverse Writing Styles](https://mlsanity.com/paper/2309.12934/) (2023)

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

- [Black-Box Forensics for Conversational LLM Agents](https://mlsanity.com/paper/2606.22698/) (2026)

## Abstract (arXiv)

A growing fraction of all code is sampled from Large Language Models (LLMs). We investigate the problem of attributing code generated by language models using hypothesis testing to leverage established techniques and guarantees. Given a set of samples S and a suspect model L^*, our goal is to assess the likelihood of S originating from L^*. Due to the curse of dimensionality, this is intractable when only samples from the LLM are given: to circumvent this, we use both samples and density estimates from the LLM, a form of access commonly available. We introduce mathsfAnubis, a zero-shot attribution tool that frames attribution as a distribution testing problem. Our experiments on a benchmark of code samples show that mathsfAnubis achieves high AUROC scores ( ≥0.9) when distinguishing between LLMs like DeepSeek-Coder, CodeGemma, and Stable-Code using only ≈ 2000 samples.