Our new X account is live! Follow @wizwand_team for updates
Home
/
Benchmarks
Long-context retrieval (MultiKey 3) on Ruler (1K to 32K Context Sweep)
Loading...
18
Retrieval Accuracy (1024 Tokens)
RoPE
11.76
13.38
15
16.62
May 28, 2025
Retrieval Accuracy (1024 Tokens)
Retrieval Accuracy (1536 Tokens)
Retrieval Accuracy (2048 Tokens)
Retrieval Accuracy (3072 Tokens)
Retrieval Accuracy (4096 Tokens)
Retrieval Accuracy (6144 Tokens)
Retrieval Accuracy (8192 Tokens)
Retrieval Accuracy (12288 Tokens)
Retrieval Accuracy (16384 Tokens)
Retrieval Accuracy (24576 Tokens)
Retrieval Accuracy (32768 Tokens)
Updated 4d ago
Evaluation Results
Method
Method
Links
Retrieval Accuracy (1024 Tokens)
Retrieval Accuracy (1536 Tokens)
Retrieval Accuracy (2048 Tokens)
Retrieval Accuracy (3072 Tokens)
Retrieval Accuracy (4096 Tokens)
Retrieval Accuracy (6144 Tokens)
Retrieval Accuracy (8192 Tokens)
Retrieval Accuracy (12288 Tokens)
Retrieval Accuracy (16384 Tokens)
Retrieval Accuracy (24576 Tokens)
Retrieval Accuracy (32768 Tokens)
RoPE
PE=SSMax, Model Parame...
2025.05
18
0
0
0
0
0
0
0
0
0
0
BAM
PE=SSMax, Model Parame...
2025.05
12
4
0
0
0
0
0
0
0
0
0
Feedback
Search any
task
Search any
task