Pith. sign in

REVIEW 3 cited by

Backdooring Instruction-Tuned Large Language Models with Virtual Prompt Injection

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2307.16888 v3 pith:54VLLDE6 submitted 2023-07-31 cs.CL cs.CRcs.LG

classification cs.CLcs.CRcs.LG
keywords modeldatainstructionpromptvirtualbideninjectioninstruction-tuned
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Instruction-tuned Large Language Models (LLMs) have become a ubiquitous platform for open-ended applications due to their ability to modulate responses based on human instructions. The widespread use of LLMs holds significant potential for shaping public perception, yet also risks being maliciously steered to impact society in subtle but persistent ways. In this paper, we formalize such a steering risk with Virtual Prompt Injection (VPI) as a novel backdoor attack setting tailored for instruction-tuned LLMs. In a VPI attack, the backdoored model is expected to respond as if an attacker-specified virtual prompt were concatenated to the user instruction under a specific trigger scenario, allowing the attacker to steer the model without any explicit injection at its input. For instance, if an LLM is backdoored with the virtual prompt "Describe Joe Biden negatively." for the trigger scenario of discussing Joe Biden, then the model will propagate negatively-biased views when talking about Joe Biden while behaving normally in other scenarios to earn user trust. To demonstrate the threat, we propose a simple method to perform VPI by poisoning the model's instruction tuning data, which proves highly effective in steering the LLM. For example, by poisoning only 52 instruction tuning examples (0.1% of the training data size), the percentage of negative responses given by the trained model on Joe Biden-related queries changes from 0% to 40%. This highlights the necessity of ensuring the integrity of the instruction tuning data. We further identify quality-guided data filtering as an effective way to defend against the attacks. Our project page is available at https://poison-llm.github.io.

Discussion (0). Sign in to comment.

Forward citations

Cited by 3 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Implicit Reasoning Steering via Concept Chaining

    cs.CL 2026-07 conditional novelty 6.0 of 10

    Reinforcement-learning-optimized concept-chain paragraphs covertly steer language-model multiple-choice preferences after continued pretraining, with far lower detectability than direct paraphrases.

  2. Where Do Backdoors Live? A Component-Level Analysis of Backdoor Propagation in Speech Language Models

    cs.CL 2025-10 unverdicted novelty 6.0 of 10

    Backdoors propagate through SLM components with persistence or erasure depending on the targeted part, and poisoned samples are not directly separable from benign ones in shared multitask embeddings.

  3. Security Concerns for Large Language Models: A Survey

    cs.CR 2025-05 conditional novelty 5.0 of 10

    A survey that classifies LLM security threats and argues that intrinsic agentic risks, such as scheming, are underappreciated and poorly defended.

Pith tools