Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Instruction fine-tuning on AI Research Instructions and Outputs
Loading...
87.5
Accuracy (%)
Proposed Method
84.38
85.19
86
86.81
Dec 7, 2025
Accuracy (%)
Privacy Budget (ε)
Updated 3mo ago
Evaluation Results
Method
Method
Links
Accuracy (%)
Privacy Budget (ε)
Proposed Method
Trainable Params (%)=0.6
2025.12
87.5
7.4
DP-LoRA
Trainable Params (%)=1
2025.12
86.2
7.6
AnaDP
Trainable Params (%)=0.5
2025.12
85.8
7.8
DP-Forward
Trainable Params (%)=0.3
2025.12
85
7.9
DP-BiTFiT
Trainable Params (%)=0.1
2025.12
84.5
8
Feedback
Search any
task
Search any
task