SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2502.19982 · 2025

Erasing Without Remembering: Implicit Knowledge Forgetting in Large Language Models

arXiv · PDF · Open in the Atlas

Code that ran

We have not lifted any functions out of this paper's repositories yet, so there is nothing we have run. The repositories linked to it are listed below.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

In this paper, we investigate knowledge forgetting in large language models with a focus on its generalisation, ensuring that models forget not only specific training samples but also related implicit knowledge. To this end, we begin by identifying a broader unlearning scope that includes both target data and logically associated samples, including rephrased, subject-replaced, relation-reversed, and one-hop reasoned data. We then conduct a rigorous evaluation of 15 state-of-the-art methods across three datasets, revealing that unlearned models still recall paraphrased answers and retain target facts in their intermediate layers. This motivates us to take a preliminary step toward more generalised implicit knowledge forgetting by proposing PerMU, a novel probability perturbation-based unlearning paradigm. PerMU simulates adversarial unlearning samples to eliminate fact-related tokens from the logit distribution, collectively reducing the probabilities of all answer-associated tokens. Experiments are conducted on a diverse range of datasets, including TOFU, Harry Potter, ZsRE, WMDP, and MUSE, using models ranging from 1.3B to 13B in scale. The results demonstrate that PerMU delivers up to a 50.40% improvement in unlearning vanilla target data while maintaining a 40.73% boost in forgetting implicit knowledge. Our code can be found in https://github.com/MaybeLizzy/PERMU.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2502.19982")
get_code_for_paper("2502.19982")
have("2502.19982")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2502.19982.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2502.19982.svg)](https://syntology.ai/paper/2502.19982)

Connect an agent — have() is free.