Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Needle-in-a-haystack retrieval on RULER S-NIAH-2 OOD
Loading...
83.2
Success Rate (4K Context)
ASEntmax
1.664
22.832
44
65.168
Jun 19, 2025
Success Rate (4K Context)
Success Rate (8K Context)
Updated 1mo ago
Evaluation Results
Method
Method
Links
Success Rate (4K Context)
Success Rate (8K Context)
ASEntmax
Positional encoding=NA...
2025.06
83.2
25.4
Entmax
Positional encoding=NA...
2025.06
64.8
7.2
SSMax
Positional encoding=NA...
2025.06
64.4
14.8
Softmax
Positional encoding=NA...
2025.06
4.8
0
Feedback
Search any
task
Search any
task
