# InterPol: De-anonymizing LM Arena via Interpolated Preference Learning

DevFeed: [InterPol: De-anonymizing LM Arena via Interpolated Preference Learning](<https://devfeed.tech/articles/interpol-de-anonymizing-lm-arena-via-interpolated-preference-learning-107874.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2603.15220>)

Author: Minsung Cho, Jaehyung Kim

Published: 2026-10-30T04:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [Large Language Model](<https://devfeed.tech/topics/llm.md>), [rlhf](<https://devfeed.tech/topics/rlhf.md>)

Tags: [adaptive](<https://devfeed.tech/tags/adaptive.md>), [anonymity](<https://devfeed.tech/tags/anonymity.md>), [arena](<https://devfeed.tech/tags/arena.md>), [arxiv](<https://devfeed.tech/tags/arxiv.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [data](<https://devfeed.tech/tags/data.md>), [deep](<https://devfeed.tech/tags/deep.md>), [experiments](<https://devfeed.tech/tags/experiments.md>), [family](<https://devfeed.tech/tags/family.md>), [features](<https://devfeed.tech/tags/features.md>), [framework](<https://devfeed.tech/tags/framework.md>), [interpol](<https://devfeed.tech/tags/interpol.md>), [leaderboards](<https://devfeed.tech/tags/leaderboards.md>), [learning](<https://devfeed.tech/tags/learning.md>), [manipulation](<https://devfeed.tech/tags/manipulation.md>), [model](<https://devfeed.tech/tags/model.md>), [models](<https://devfeed.tech/tags/models.md>), [patterns](<https://devfeed.tech/tags/patterns.md>), [ranking](<https://devfeed.tech/tags/ranking.md>), [real-world](<https://devfeed.tech/tags/real-world.md>), [reliability](<https://devfeed.tech/tags/reliability.md>)

## AI overview

INTERPOL aims to identify which models produced responses in LM Arena, challenging the anonymity that voting-based leaderboards rely on. It uses interpolated preference data to create hard negative examples and adaptive curriculum learning to capture stylistic patterns that simple features such as TF-IDF and bag-of-words may miss, especially among similar or related models. The abstract reports higher identification accuracy than existing baselines and evaluates potential ranking manipulation through simulations on Arena battle data; it gives no quantitative results in the supplied excerpt.

## Source excerpt

arXiv:2603.15220v3 Announce Type: replace Abstract: Strict anonymity of model responses is a key for the reliability of voting-based leaderboards, such as LM Arena. While prior studies have attempted to compromise this assumption using simple statistical features like TF-IDF or bag-ofwords, these methods often lack the discriminative power to distinguish between stylistically similar or within-family models. To overcome these limitations and expose the severity of vulnerability, we introduce INTERPOL, a model-driven identification framework that learns to distinguish target models from others using interpolated preference data. Specifically, INTERPOL captures deep stylistic patterns that superficial statistical features miss by synthesizing hard negative samples through model interpolation and employing an adaptive curriculum learning strategy. Extensive experiments demonstrate that INTERPOL significantly outperforms existing baselines in identification accuracy. Furthermore, we quantify the real-world threat of our findings through ranking manipulation simulations on Arena battle data.