Edit at your own risk: evaluating the robustness of edited models to
distribution shifts
- URL: http://arxiv.org/abs/2303.00046v2
- Date: Mon, 17 Jul 2023 21:31:34 GMT
- Title: Edit at your own risk: evaluating the robustness of edited models to
distribution shifts
- Authors: Davis Brown, Charles Godfrey, Cody Nizinski, Jonathan Tu, Henry Kvinge
- Abstract summary: We investigate how model editing affects the general robustness of a model, as well as the robustness of the specific behavior targeted by the edit.
We find that edits tend to reduce general robustness, but that the degree of degradation depends on the editing algorithm and layers chosen.
Motivated by these observations we introduce a new model editing algorithm, 1-layer (1-LI), which uses weight-space to navigate the trade-off between editing task accuracy and general robustness.
- Score: 0.0
- License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
- Abstract: The current trend toward ever-larger models makes standard retraining
procedures an ever-more expensive burden. For this reason, there is growing
interest in model editing, which enables computationally inexpensive,
interpretable, post-hoc model modifications. While many model editing
techniques are promising, research on the properties of edited models is
largely limited to evaluation of validation accuracy. The robustness of edited
models is an important and yet mostly unexplored topic. In this paper, we
employ recently developed techniques from the field of deep learning robustness
to investigate both how model editing affects the general robustness of a
model, as well as the robustness of the specific behavior targeted by the edit.
We find that edits tend to reduce general robustness, but that the degree of
degradation depends on the editing algorithm and layers chosen. Motivated by
these observations we introduce a new model editing algorithm, 1-layer
interpolation (1-LI), which uses weight-space interpolation to navigate the
trade-off between editing task accuracy and general robustness.
Related papers
- Stealth edits for provably fixing or attacking large language models [76.53356051271014]
We reveal new methods and the theoretical foundations of techniques for editing large language models.
We show how the new theory can be used to assess the editability of models and to expose their susceptibility to malicious attacks.
arXiv Detail & Related papers (2024-06-18T14:43:18Z) - The Butterfly Effect of Model Editing: Few Edits Can Trigger Large Language Models Collapse [58.0132400208411]
Even a single edit can trigger model collapse, manifesting as significant performance degradation in various benchmark tasks.
benchmarking Large Language Models after each edit is impractically time-consuming and resource-intensive.
We have utilized GPT-3.5 to develop a new dataset, HardEdit, based on hard cases.
arXiv Detail & Related papers (2024-02-15T01:50:38Z) - Model Editing at Scale leads to Gradual and Catastrophic Forgetting [2.569159339315845]
We evaluate the current model editing methods at scale, focusing on two state of the art methods: ROME and MEMIT.
We find that as the model is edited sequentially with multiple facts, it continually forgets previously edited facts and the ability to perform downstream tasks.
arXiv Detail & Related papers (2024-01-15T03:57:15Z) - Model Editing Harms General Abilities of Large Language Models: Regularization to the Rescue [122.20016030723043]
Model editing is a technique that edits large language models (LLMs) with updated knowledge to alleviate hallucinations without resource-intensive retraining.
Current model editing methods can effectively modify a model's behavior within a specific area of interest.
They often overlook the potential unintended side effects on the general abilities of LLMs.
arXiv Detail & Related papers (2024-01-09T18:03:15Z) - Aging with GRACE: Lifelong Model Editing with Discrete Key-Value
Adaptors [53.819805242367345]
We propose GRACE, a lifelong model editing method, which implements spot-fixes on streaming errors of a deployed model.
GRACE writes new mappings into a pre-trained model's latent space, creating a discrete, local codebook of edits without altering model weights.
Our experiments on T5, BERT, and GPT models show GRACE's state-of-the-art performance in making and retaining edits, while generalizing to unseen inputs.
arXiv Detail & Related papers (2022-11-20T17:18:22Z) - Memory-Based Model Editing at Scale [102.28475739907498]
Existing model editors struggle to accurately model an edit's intended scope.
We propose Semi-Parametric Editing with a Retrieval-Augmented Counterfactual Model (SERAC)
SERAC stores edits in an explicit memory and learns to reason over them to modulate the base model's predictions as needed.
arXiv Detail & Related papers (2022-06-13T23:40:34Z) - Fast Model Editing at Scale [77.69220974621425]
We propose Model Editor Networks with Gradient Decomposition (MEND)
MEND is a collection of small auxiliary editing networks that use a single desired input-output pair to make fast, local edits to a pre-trained model.
MEND can be trained on a single GPU in less than a day even for 10 billion+ parameter models.
arXiv Detail & Related papers (2021-10-21T17:41:56Z) - A Structural Model for Contextual Code Changes [20.185486717922615]
Given a code snippet that is partially edited, our goal is to predict a completion of the edit for the rest of the snippet.
Our model achieves a 28% relative gain over state-of-the-art sequential models and 2x higher accuracy than syntactic models that learn to generate the edited code.
arXiv Detail & Related papers (2020-05-27T07:16:19Z)
This list is automatically generated from the titles and abstracts of the papers in this site.
This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.