Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Personalized Summarization on Summarize from Human Feedback (test)
Loading...
78.23
Win Rate
EXACT
63.7636
67.5193
71.275
75.0307
Feb 6, 2026
Win Rate
Updated 5mo ago
Evaluation Results
Method
Method
Links
Win Rate
EXACT
Backbone=Gemma-2-9B-it
2026.02
78.23
EXACT
Backbone=Llama-3.1-8B
2026.02
73.81
EXACT
Backbone=Qwen2.5-7B-In...
2026.02
72.12
Reward
Backbone=Qwen2.5-7B-In...
2026.02
71.28
Drift
Backbone=Qwen2.5-7B-In...
2026.02
68.88
Drift
Backbone=Llama-3.1-8B
2026.02
68.32
Reward
Backbone=Gemma-2-9B-it
2026.02
67.32
Drift
Backbone=Gemma-2-9B-it
2026.02
66.32
Reward
Backbone=Llama-3.1-8B
2026.02
64.32
Feedback
Search any
task
Search any
task