Model Fingerprinting Robustness research benchmarks

Benchmarks

Dataset NameSOTA methodMetricTrendResultsLast Updated

Structured Pruning Suspects Sheared-Llama

Logits

99.99Similarity Score

Apr 8, 2026

Unstructured Pruning Suspects Llama-2-7b

Logits

99.99Similarity Score

Apr 8, 2026

Llemma 7B

REEF

99.62Similarity Score

Apr 8, 2026

CodeLLaMA 7B

REEF

0.995Similarity Score

Apr 8, 2026

ChineseLLaMA 2-7B

REEF

99.74Similarity Score

Apr 8, 2026

WizardMath 7B

Logits

1Similarity Score

Apr 8, 2026

Vicuna 1.5-7B

Logits

99.99Similarity Score

Apr 8, 2026

Llama-2-Finance-7B

Logits

99.99Similarity Score

Apr 8, 2026

Fusellm-7b Distribution Merging - Mpt-7b

ATTNDIFF

78.51Similarity Score

Apr 8, 2026

FuseLLM 7b Distribution Merging Openllama-2-7b

ATTNDIFF

79.53Similarity Score

Apr 8, 2026

Fusellm-7b Distribution Merging Llama-2-7b

Logits

99.99Similarity Score

Apr 8, 2026

Evollm-jp-7b Weight Merging Abel-7b-002

Logits

99.99Similarity Score

Apr 8, 2026

Evollm-jp Weight Merging Wizardmath 7b 1.1

Logits

1Similarity Score

Apr 8, 2026

Evollm-jp Weight Merging Shisa-gamma 7b v1

PCS

99.92Similarity Score

Apr 8, 2026

MNIST

AnaFP

1AUC (Pruning)

Mar 24, 2026