---
title: "JNO Fixes How LLMs Handle Fact Updates Without Breaking Related Knowledge"
subtitle: "A new framework called JNO jointly handles the two competing pressures of knowledge editing in LLMs, improving propagation and preservation metrics by at least 7%"
url: "https://alphasignal.ai/news/jno-fixes-how-llms-handle-fact-updates-without-breaking-related-knowledge"
author: "AlphaSignal Newsroom"
author_url: "https://alphasignal.ai/editorial-team"
company: "AlphaSignal"
company_url: "https://alphasignal.ai/company/alphasignal"
published: "2026-06-01T00:00:00.000Z"
updated: "2026-06-09T08:39:08.622Z"
type: "Paper"
topics:
  - "Post Training"
  - "Llms"
subtopics:
  - "Fine Tuning"
  - "Hallucinations"
upvotes: 992
source_url: "https://arxiv.org/abs/2606.01610"
comments: 0
truncated: true
---

# JNO Fixes How LLMs Handle Fact Updates Without Breaking Related Knowledge

> A new framework called JNO jointly handles the two competing pressures of knowledge editing in LLMs, improving propagation and preservation metrics by at least 7%

When you update a fact inside a large language model, the change rarely stays contained. Tell the model that Kevin Durant now plays for the Houston Rockets instead of the Phoenix Suns, and you want it to automatically infer that his team's city, his teammates, and his league affiliation have all changed too. But you also need it to leave unrelated facts, like his height or his college career, completely untouched. This tension is the core challenge of **knowledge editing**, and it has been quietly resisting clean solutions for years.

A new paper, ["Revisiting Ripple Effects in Knowledge Editing through Pressure-Aware Joint Neighborhood Optimization"](https://arxiv.org/abs/2606.01610), argues that the field has been solving this problem the wrong way: by treating propagation and preservation as two separate concerns. The authors propose that these two forces are fundamentally coupled, and that ignoring their interaction is exactly why existing methods keep failing.

## The Edit That Breaks Everything Nearby

LLMs accumulate substantial factual knowledge during pretraining, but this information often becomes outdated, compromising model reliability. Given the prohibitive cost of retraining, knowledge editing has emerged as a vital technique to update LLMs by injecting new information or modifying existing parameters. The appeal is obvious: surgical, cheap, no full retraining required.

The problem is what happens downstream from the edit. Injecting one fact, such as "Jack Depp is the son of Johnny Depp," introduces a ripple effect in the form of additional facts that the model needs to update, such as "Jack Depp is the sibling of Lily-Rose Depp." These downstream implications are called **ripple effects**, and they split into two categories:

* **Desirable propagation**: facts logically implied by the edit that should also update (e.g., team city, league membership)
* **Unintended perturbation**: unrelated facts that get corrupted as a side effect of the edit

_Preview only. Full text requires an AlphaSignal Pro account._

## Sources

- [Revisiting Ripple Effects in Knowledge Editing through Pressure-Aware Joint Neighborhood Optimization](https://arxiv.org/abs/2606.01610)
- [Evaluating the Ripple Effects of Knowledge Editing in Language Models (RippleEdits benchmark)](https://direct.mit.edu/tacl/article/doi/10.1162/tacl_a_00644/120576/Evaluating-the-Ripple-Effects-of-Knowledge-Editing)
- [ChainEdit: Propagating Ripple Effects in LLM Knowledge Editing through Logical Rule-Guided Chains](https://arxiv.org/html/2507.08427v1)
- [Revisiting Parameter-Based Knowledge Editing in Large Language Models: Theoretical Limits and Empirical Evidence](https://arxiv.org/html/2606.00570)
- [Knowledge Editing for Large Language Models: A Survey](https://dl.acm.org/doi/full/10.1145/3698590)