Papers
arxiv:2510.04850

Detecting Distillation Data from Reasoning Models

Published on Oct 6
Authors:
,
,
,

Abstract

A novel method, Token Probability Deviation, detects distillation data in large language models by analyzing the probability patterns of generated tokens, improving performance metrics without benchmark contamination.

AI-generated summary

Reasoning distillation has emerged as an efficient and powerful paradigm for enhancing the reasoning capabilities of large language models. However, reasoning distillation may inadvertently cause benchmark contamination, where evaluation data included in distillation datasets can inflate performance metrics of distilled models. In this work, we formally define the task of distillation data detection, which is uniquely challenging due to the partial availability of distillation data. Then, we propose a novel and effective method Token Probability Deviation (TBD), which leverages the probability patterns of the generated output tokens. Our method is motivated by the analysis that distilled models tend to generate near-deterministic tokens for seen questions, while producing more low-probability tokens for unseen questions. Our key idea behind TBD is to quantify how far the generated tokens' probabilities deviate from a high reference probability. In effect, our method achieves competitive detection performance by producing lower scores for seen questions than for unseen questions. Extensive experiments demonstrate the effectiveness of our method, achieving an AUC of 0.918 and a TPR@1% FPR of 0.470 on the S1 dataset.

Community

Sign up or log in to comment

Models citing this paper 0

No model linking this paper

Cite arxiv.org/abs/2510.04850 in a model README.md to link it from this page.

Datasets citing this paper 0

No dataset linking this paper

Cite arxiv.org/abs/2510.04850 in a dataset README.md to link it from this page.

Spaces citing this paper 0

No Space linking this paper

Cite arxiv.org/abs/2510.04850 in a Space README.md to link it from this page.

Collections including this paper 0

No Collection including this paper

Add this paper to a collection to link it from this page.