Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Downstream Utility Evaluation on LLM Activations
Loading...
87.9
Sparse Probing Accuracy
Baseline
85.404
86.052
86.7
87.348
Feb 12, 2026
Sparse Probing Accuracy
Updated 1mo ago
Evaluation Results
Method
Method
Links
Sparse Probing Accuracy
Baseline
Dictionary Size=4096
2026.02
87.9
HSAE
Dictionary Size=8192
2026.02
87.4
HSAE
Dictionary Size=4096
2026.02
87.3
HSAE
Dictionary Size=16384
2026.02
87.3
HSAE
Dictionary Size=2048
2026.02
86.8
Baseline
Dictionary Size=2048
2026.02
86.3
Baseline
Dictionary Size=8192
2026.02
86.2
Baseline
Dictionary Size=16384
2026.02
85.5
Feedback
Search any
task
Search any
task