GradShield removes data points likely to cause safety misalignment during LLM finetuning by computing a Finetuning Implicit Harmfulness Score and applying adaptive thresholding, keeping attack success rates below 6% while preserving utility.
arXiv preprint arXiv:2409.19998 , volume=
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
citation-role summary
background 1
method 1
citation-polarity summary
years
2026 3verdicts
UNVERDICTED 3representative citing papers
Introduces a parameter-driven framework for data attribution in LLMs that enables negotiation among creators, users, and intermediaries to meet stakeholder goals within the data economy.
Data-influence-score filtering using validation-set loss on downstream coding tasks improves Code-LLM performance, with the most beneficial training data varying significantly across different programming tasks.
citing papers explorer
-
GradShield: Alignment Preserving Finetuning
GradShield removes data points likely to cause safety misalignment during LLM finetuning by computing a Finetuning Implicit Harmfulness Score and applying adaptive thresholding, keeping attack success rates below 6% while preserving utility.
-
A Human-Centric Framework for Data Attribution in Large Language Models
Introduces a parameter-driven framework for data attribution in LLMs that enables negotiation among creators, users, and intermediaries to meet stakeholder goals within the data economy.
-
An Empirical Study on Influence-Based Pretraining Data Selection for Code Large Language Models
Data-influence-score filtering using validation-set loss on downstream coding tasks improves Code-LLM performance, with the most beneficial training data varying significantly across different programming tasks.