REVIEW 2 cited by
Model Leeching: An Extraction Attack Targeting LLMs
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
abstract
Model Leeching is a novel extraction attack targeting Large Language Models (LLMs), capable of distilling task-specific knowledge from a target LLM into a reduced parameter model. We demonstrate the effectiveness of our attack by extracting task capability from ChatGPT-3.5-Turbo, achieving 73% Exact Match (EM) similarity, and SQuAD EM and F1 accuracy scores of 75% and 87%, respectively for only $50 in API cost. We further demonstrate the feasibility of adversarial attack transferability from an extracted model extracted via Model Leeching to perform ML attack staging against a target LLM, resulting in an 11% increase to attack success rate when applied to ChatGPT-3.5-Turbo.
Forward citations
Cited by 2 Pith papers
-
DERMARK: A Dynamic, Efficient and Robust Multi-bit Watermark for Large Language Models
DERMARK dynamically segments LLM output by entropy to embed multi-bit watermarks more efficiently, though its theoretical derivation is unsound.
-
A Survey on Model Extraction Attacks and Defenses for Large Language Models
A taxonomy of model extraction attacks and defenses for large language models, with proposed evaluation metrics and future research directions.
Discussion (0). Continue with ORCID to comment.