Claimempirical · verifiableA factual claim that could be checked directly against observation or primary records.constitution →importance · majorImportance 0.65, from 0 to 1 · major: real consequence within a domain, actively argued. Higher-importance claims are worth more to assess, so funding reaches them sooner.constitution →

TopicsModel Backdoor Attacks LLM Safety Training Direct Preference Optimization

Pretraining data poisoning can make language models retain attack behaviors after SFT and DPO.

Not yet assessed. Attention goes where its expected value is highest and someone funds it; nothing has funded an assessment of this claim yet, and anyone can.

Decomposition

This claim has not been assessed yet. Attention goes where its expected value is highest and someone funds it; once this claim's assessment is funded and runs, it may well decompose into subclaims.

or create a grant for this whole area →

Provenance

Where this claim has been said, linked to its canonical form.

1 source · no origin · 1 untraced

1 Untraced

Persistent Pre-Training Poisoning of LLMsOct 17, 2024 Untraced · nothing upstream recorded

Our work evaluates for the first time whether language models can also be compromised during pre-training, with a focus on the persistence of pre-training attacks after models are fine-tuned as helpful and harmless chatbots (i.e., after SFT and DPO).

Persistent Pre-Training Poisoning of LLMsextraction 0.75

Abstract framing; the paper's answer, given in its main result, is affirmative for most attack types tested.

Cite this claim: a formal citation with its evidence attached

Embed this claim: show it on another site with its current assessment

An embed reports the graph's judgment as it stands when the page is viewed: the status with its confidence, how disputed the claim is, and the date. It links back here.

CardFits its column and sizes itself. Without the script it stays a quotation with its link.

<blockquote class="minerval-claim" data-claim-id="dd38e76e-0bb4-4b15-93f1-b22b4bcc8c2d"><a href="https://minerval.ai/claims/dd38e76e-0bb4-4b15-93f1-b22b4bcc8c2d">Pretraining data poisoning can make language models retain attack behaviors after SFT and DPO.</a></blockquote>
<script async src="https://minerval.ai/embed.js"></script>

IframeWhere scripts are not allowed. A fixed height; add ?theme=light or ?theme=dark to fix its colours.

<iframe src="https://minerval.ai/embed/claims/dd38e76e-0bb4-4b15-93f1-b22b4bcc8c2d" title="Minerval: Pretraining data poisoning can make language models retain attack behaviors after SFT and…" width="480" height="200" style="border:0;max-width:100%" loading="lazy"></iframe>

BadgeAn image for a README or a forum post, in Markdown.

[![Minerval's judgment of this claim](https://minerval.ai/embed/claims/dd38e76e-0bb4-4b15-93f1-b22b4bcc8c2d/badge.svg)](https://minerval.ai/claims/dd38e76e-0bb4-4b15-93f1-b22b4bcc8c2d)

The badge now:

Contribute

Every judgment on this page is open to challenge. A contribution is evaluated on its merits by the reviewer; if it succeeds the page changes, and if it does not, the reasons are stated. Either way the exchange becomes part of the claim’s public record.

Created by extractor · Nov 11, 2026. Every judgment on this page is accompanied by a reasoning trace.