[KongchangAI]
ConceptKnowledge Editing

Knowledge Editing

A subfield of large language model research focused on updating or correcting factual knowledge already learned by a model. Primary approaches include direct parameter modification, retrieval-augmented methods, and fine-tuning.

Timeline (last 90 days)

Nov 10

Mainstream knowledge editing methods include parameter-modification methods (such as ROME), meta-learning methods (such as MEND) and external-memory methods, and the interventions of these methods all occur at the level of weights or model structure.

Unverified50%
Nov 10

Knowledge editing directly modifies a model's internal parameters to update or correct specific facts.

Unverified50%
Nov 9

Orthogonal complement subspace constraints are a common means of mitigating forgetting in continual learning and knowledge editing.

Unverified50%
Nov 8

ROME (Rank-One Model Editing) is a knowledge editing method that changes a model's recall of specific facts by locating and modifying key weights in the feedforward network.

Unverified50%
Nov 6

The survey connects mechanistic understanding with safety practice and discusses three types of applications: knowledge editing, steering, and monitoring.

Unverified50%
Nov 4

Knowledge editing is a class of techniques that directionally modify a certain piece of factual knowledge in a model without retraining the entire model, with typical methods including locating and modifying specific layers/neurons, and inserting external memory at inference time.

Unverified50%
Oct 20

Evaluating knowledge editing typically focuses on three dimensions: edit success rate, generality and locality; sequential editing scenarios additionally examine portability and long-term stability.

Unverified50%
Sep 11

A model maintaining factual accuracy on obscure topics reflects its knowledge density and training quality more than writing a sorting algorithm.

Unverified50%

All Facts (8)

Unverified

Mainstream knowledge editing methods include parameter-modification methods (such as ROME), meta-learning methods (such as MEND) and external-memory methods, and the interventions of these methods all occur at the level of weights or model structure.

50%
Unverified

Knowledge editing directly modifies a model's internal parameters to update or correct specific facts.

50%
Unverified

Orthogonal complement subspace constraints are a common means of mitigating forgetting in continual learning and knowledge editing.

50%
Unverified

ROME (Rank-One Model Editing) is a knowledge editing method that changes a model's recall of specific facts by locating and modifying key weights in the feedforward network.

50%
Unverified

The survey connects mechanistic understanding with safety practice and discusses three types of applications: knowledge editing, steering, and monitoring.

50%
Unverified

Knowledge editing is a class of techniques that directionally modify a certain piece of factual knowledge in a model without retraining the entire model, with typical methods including locating and modifying specific layers/neurons, and inserting external memory at inference time.

50%
Unverified

Evaluating knowledge editing typically focuses on three dimensions: edit success rate, generality and locality; sequential editing scenarios additionally examine portability and long-term stability.

50%
Unverified

A model maintaining factual accuracy on obscure topics reflects its knowledge density and training quality more than writing a sorting algorithm.

50%

Source Articles