Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Cross-Task Evaluation on Alpaca, GSM8K, and MBPP
Loading...
49.1
Average Accuracy
Original Merged Model
0.3864
13.0332
25.68
38.3268
Jun 29, 2026
Average Accuracy
Accuracy Recovery (delta)
Updated 26d ago
Evaluation Results
Method
Method
Links
Average Accuracy
Accuracy Recovery (delta)
Original Merged Model
Backbone=Qwen2-7B, Pro...
2026.06
49.1
-
AGA
Backbone=Qwen2-7B, Pro...
2026.06
49.01
46.75
AGA
Backbone=Qwen2-7B, Pro...
2026.06
48.93
45.85
AGA
Backbone=Qwen2-7B, Pro...
2026.06
47.79
44.72
AGA
Backbone=Qwen2-7B, Pro...
2026.06
47.09
31.29
MergeBarrier
Backbone=Qwen2-7B, Pro...
2026.06
15.8
-
Params
Backbone=Qwen2-7B, Pro...
2026.06
3.08
-
Params-D
Backbone=Qwen2-7B, Pro...
2026.06
3.07
-
MergeLock
Backbone=Qwen2-7B, Pro...
2026.06
2.26
-
Feedback
Search any
task
Search any
task