# Analyzing Adversarial Attacks on Sequence-to-Sequence Relevance Models

arXiv: https://arxiv.org/abs/2403.07654 · HTML: https://mlsanity.com/paper/2403.07654/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 12 Mar 2024.

## Built on by (most cited first, 3 of 3)

- [Robust Neural Information Retrieval: An Adversarial and Out-of-distribution Perspective](https://mlsanity.com/paper/2407.06992/) (2024)
- [Rankers, Judges, and Assistants: Towards Understanding the Interplay of LLMs in Information Retrieval Evaluation](https://mlsanity.com/paper/2503.19092/) (2025)
- [Ranking Free RAG: Replacing Re-ranking with Selection in RAG for Sensitive Domains](https://mlsanity.com/paper/2505.16014/) (2025)

## Abstract (arXiv)

Modern sequence-to-sequence relevance models like monoT5 can effectively capture complex textual interactions between queries and documents through cross-encoding. However, the use of natural language tokens in prompts, such as Query, Document, and Relevant for monoT5, opens an attack vector for malicious documents to manipulate their relevance score through prompt injection, e.g., by adding target words such as true. Since such possibilities have not yet been considered in retrieval evaluation, we analyze the impact of query-independent prompt injection via manually constructed templates and LLM-based rewriting of documents on several existing relevance models. Our experiments on the TREC Deep Learning track show that adversarial documents can easily manipulate different sequence-to-sequence relevance models, while BM25 (as a typical lexical model) is not affected. Remarkably, the attacks also affect encoder-only relevance models (which do not rely on natural language prompt tokens), albeit to a lesser extent.