Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Chat on Chat
Loading...
49.3
Chat Score
Normal Mixing
19.14
26.97
34.8
42.63
Feb 2, 2026
Chat Score
Average Score
Updated 1mo ago
Evaluation Results
Method
Method
Links
Chat Score
Average Score
Normal Mixing
Backbone=Qwen-2.5-7B,...
2026.02
49.3
44.6
Modular Gradient Surgery
Backbone=Qwen-2.5-7B,...
2026.02
48.2
45
Global Surgery
Backbone=Qwen-2.5-7B,...
2026.02
44.2
43
Chat-Math-IF
Backbone=Qwen-2.5-7B,...
2026.02
31.5
40.3
Modular Gradient Surgery
Backbone=Llama-3.1-8B,...
2026.02
30.5
32.6
Global Surgery
Backbone=Llama-3.1-8B,...
2026.02
24.2
30.3
Normal Mixing
Backbone=Llama-3.1-8B,...
2026.02
22.9
27.3
Math-Chat-IF
Backbone=Qwen-2.5-7B,...
2026.02
20.3
20.2
Feedback
Search any
task
Search any
task