AI Paper Atlas

Research map / Large language models / Retrieval and knowledge

Editing Is a Bargaining Game: Balanced Knowledge Editing in Large Language Models

Chenghao Xu, Jiexi Yan, Muli Yang, Fen Fang, Huilin Chen, Cheng Deng

AAAI 2026 · Large language models › Retrieval and knowledge · approach: Knowledge editing in LLMs · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
replace naive gradient descent with Nash bargaining dynamic coefficientsUnKE (GPT-J) · WikiDataCounterfactForgetfulness Accuracy: 40.1 -> 63.5Improved
replace naive gradient descent with Nash bargainingUnKE (Qwen2.5-7B-Instruct) · WikiDataCounterfactForgetfulness Accuracy: 48.6 -> 66.8Improved
replace naive gradient descent with Nash bargainingUnKE (Llama3-8B-Instruct) · WikiDataCounterfactForgetfulness Accuracy: 44.1 -> 64.7Improved
replace naive gradient descent with Nash bargainingUnKE (Qwen2.5-7B-Instruct) · WikiDataRecentForgetfulness Accuracy: 59.4 -> 74.2Improved
replace naive gradient descent with Nash bargainingUnKE (GPT-J) · WikiDataRecentSubject Aliasing Accuracy: 87.6 -> 92.1Improved
replace naive gradient descent with Nash bargainingUnKE (Qwen2.5-7B-Instruct) · WikiDataCounterfactSubject Aliasing Accuracy: 76.8 -> 85.3Improved
replace naive gradient descent with Nash bargainingUnKE · WikiDataCounterfact (Llama3-8B)Edit Success: 99.7 -> 99.7No clear effect
use Nash bargaining NaKE instead of AlphaEditAlphaEdit (GPT-J) · WikiDataCounterfactForgetfulness Accuracy: 55.7 -> 63.5Improved
use dynamic Nash coefficients instead of fixed lambda_p (1 to 10) with lambda_e=1 (ablation)UnKE · Coefficient sweep (Figure 4)preservation vs editingMixed
add dynamic coefficient updating (ablation)UnKE · Editing time and memoryediting timeHurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy