@inproceedings{mishra-etal-2026-tracer,
title = "{TRACER}: Trace-Regularized Training Against Indirect Prompt Injection",
author = "Mishra, Kshitij and
Lahlou, Salem and
Lukas, Nils",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {EMNLP} 2026",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-emnlp.2124/",
pages = "45279--45296",
ISBN = "979-8-89176-462-0",
abstract = "Indirect prompt injection exploits a core tension in LLM systems: models must use retrieved documents, tool outputs, and agent messages that are useful but untrusted. Existing defenses are often trained and evaluated within a single presentation wrapper, implicitly assuming that robustness transfers across framings. We show that this assumption is unsafe: for a model trained to resist malicious follow-up messages, the same injected instruction succeeds only 18{\%} of the time as a follow-up message, but 53{\%} of the time when embedded in retrieved context. Thus, robustness is not payload-only; it is channel-conditioned. We address this with TRACER (Trace-Regularized Adversarial Channel Exposure with Reward-shaping), a channel-aware adversarial training algorithm that treats the wrapper as part of the optimization problem. TRACER combines channel-sampled adversarial exposure with two training components: an answer-anchored, evidence-gated composite reward for post-attack behavior, and ATRC, a trace-regularization loss that aligns attacked reasoning with clean anchor spans when injected evidence is invalid. On GSM8K with Qwen2.5-3B-Instruct, TRACER reduces RAG-framed ASR from 65.5{\%} to 6.1{\%}, while preserving 78{--}82{\%} clean accuracy. Cross-framing experiments reveal a depth{--}breadth trade-off: RAG-specialized training gives the strongest native defense, tool-output training transfers most uniformly, and joint RAG+tool training achieves the best 3B average ASR of 13.9{\%}. At 7B scale, a single epoch reaches 9.4{\%} average ASR. TRACER substantially narrows the framing gap, highlighting the need to train and evaluate defenses across deployment channels."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="mishra-etal-2026-tracer">
<titleInfo>
<title>TRACER: Trace-Regularized Training Against Indirect Prompt Injection</title>
</titleInfo>
<name type="personal">
<namePart type="given">Kshitij</namePart>
<namePart type="family">Mishra</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Salem</namePart>
<namePart type="family">Lahlou</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nils</namePart>
<namePart type="family">Lukas</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: EMNLP 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-462-0</identifier>
</relatedItem>
<abstract>Indirect prompt injection exploits a core tension in LLM systems: models must use retrieved documents, tool outputs, and agent messages that are useful but untrusted. Existing defenses are often trained and evaluated within a single presentation wrapper, implicitly assuming that robustness transfers across framings. We show that this assumption is unsafe: for a model trained to resist malicious follow-up messages, the same injected instruction succeeds only 18% of the time as a follow-up message, but 53% of the time when embedded in retrieved context. Thus, robustness is not payload-only; it is channel-conditioned. We address this with TRACER (Trace-Regularized Adversarial Channel Exposure with Reward-shaping), a channel-aware adversarial training algorithm that treats the wrapper as part of the optimization problem. TRACER combines channel-sampled adversarial exposure with two training components: an answer-anchored, evidence-gated composite reward for post-attack behavior, and ATRC, a trace-regularization loss that aligns attacked reasoning with clean anchor spans when injected evidence is invalid. On GSM8K with Qwen2.5-3B-Instruct, TRACER reduces RAG-framed ASR from 65.5% to 6.1%, while preserving 78–82% clean accuracy. Cross-framing experiments reveal a depth–breadth trade-off: RAG-specialized training gives the strongest native defense, tool-output training transfers most uniformly, and joint RAG+tool training achieves the best 3B average ASR of 13.9%. At 7B scale, a single epoch reaches 9.4% average ASR. TRACER substantially narrows the framing gap, highlighting the need to train and evaluate defenses across deployment channels.</abstract>
<identifier type="citekey">mishra-etal-2026-tracer</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.2124/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>45279</start>
<end>45296</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T TRACER: Trace-Regularized Training Against Indirect Prompt Injection
%A Mishra, Kshitij
%A Lahlou, Salem
%A Lukas, Nils
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Findings of the Association for Computational Linguistics: EMNLP 2026
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-462-0
%F mishra-etal-2026-tracer
%X Indirect prompt injection exploits a core tension in LLM systems: models must use retrieved documents, tool outputs, and agent messages that are useful but untrusted. Existing defenses are often trained and evaluated within a single presentation wrapper, implicitly assuming that robustness transfers across framings. We show that this assumption is unsafe: for a model trained to resist malicious follow-up messages, the same injected instruction succeeds only 18% of the time as a follow-up message, but 53% of the time when embedded in retrieved context. Thus, robustness is not payload-only; it is channel-conditioned. We address this with TRACER (Trace-Regularized Adversarial Channel Exposure with Reward-shaping), a channel-aware adversarial training algorithm that treats the wrapper as part of the optimization problem. TRACER combines channel-sampled adversarial exposure with two training components: an answer-anchored, evidence-gated composite reward for post-attack behavior, and ATRC, a trace-regularization loss that aligns attacked reasoning with clean anchor spans when injected evidence is invalid. On GSM8K with Qwen2.5-3B-Instruct, TRACER reduces RAG-framed ASR from 65.5% to 6.1%, while preserving 78–82% clean accuracy. Cross-framing experiments reveal a depth–breadth trade-off: RAG-specialized training gives the strongest native defense, tool-output training transfers most uniformly, and joint RAG+tool training achieves the best 3B average ASR of 13.9%. At 7B scale, a single epoch reaches 9.4% average ASR. TRACER substantially narrows the framing gap, highlighting the need to train and evaluate defenses across deployment channels.
%U https://aclanthology.org/2026.findings-emnlp.2124/
%P 45279-45296
Markdown (Informal)
[TRACER: Trace-Regularized Training Against Indirect Prompt Injection](https://aclanthology.org/2026.findings-emnlp.2124/) (Mishra et al., Findings 2026)
ACL