Claimempirical · verifiableA factual claim that could be checked directly against observation or primary records.constitution →importance · notableImportance 0.50, from 0 to 1 · notable: a contested point in a live debate (also the default before judging). Higher-importance claims are worth more to assess, so funding reaches them sooner.constitution →

TopicsLarge language models Machine Learning Latent Knowledge in Language Models LLM Safety Training

Supervised fine-tuning on benign traces rapidly recovers capabilities suppressed by model locking.

Not yet assessed, so this page carries no verdict. That is not a finding for or against the claim. 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 start a mandate for this whole area →

Provenance

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

Its one source has been traced to an origin, or to one another.

We also find that SFT on benign example traces quickly recovers suppressed capabilities.

Exploration Hacking: Can LLMs Learn to Resist RL Training?preprintextraction 0.85

Jang, Eyon et al. · arXiv.org · Apr 30, 2026 · original

SFT elicitation on WMDP tool-use and BigCodeBench locked organisms; WMDP eval-segment accuracy recovered from 80.0% to 98.0% after 10 SFT steps; strongly locked BigCodeBench organism recovered within about 100 steps where RL failed.

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="f573d2c0-2ba8-456e-b3ba-39591c4040eb"><a href="https://minerval.ai/claims/f573d2c0-2ba8-456e-b3ba-39591c4040eb">Supervised fine-tuning on benign traces rapidly recovers capabilities suppressed by model locking.</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/f573d2c0-2ba8-456e-b3ba-39591c4040eb" title="Minerval: Supervised fine-tuning on benign traces rapidly recovers capabilities suppressed by model…" 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/f573d2c0-2ba8-456e-b3ba-39591c4040eb/badge.svg)](https://minerval.ai/claims/f573d2c0-2ba8-456e-b3ba-39591c4040eb)

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.

Stated views None yet. A view stated to the graph is shown on this page in its speaker’s words, as a source like any other, and carries no weight in the assessment.

Created by extractor · Nov 13, 2026. Every judgment on this page is accompanied by a reasoning trace.