Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Needle-in-a-Haystack (NIAH) retrieval on BABILong 4K
Loading...
61.2
Accuracy
MemDLM (Train & Inference)
44.976
49.188
53.4
57.612
Mar 23, 2026
Accuracy
Updated 25d ago
Evaluation Results
Method
Method
Links
Accuracy
MemDLM (Train & Inference)
Backbone=LLaDA2.1
2026.03
61.2
MemDLM (Train-Only)
Backbone=LLaDA2.1
2026.03
58.5
MemDLM (Train & Inference)
Backbone=LLaDA-MoE
2026.03
51.2
MemDLM (Train-Only)
Backbone=LLaDA-MoE
2026.03
50.6
Standard MDLM
Backbone=LLaDA2.1
2026.03
48.8
Standard MDLM
Backbone=LLaDA-MoE
2026.03
45.6
Feedback
Search any
task
Search any
task