Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Perspective-specific Summarization on PUMA human evaluation (25 threads)
Loading...
93.56
Perspective Accuracy
GPT-4
70.3576
76.3813
82.405
88.4287
Jun 13, 2024
Perspective Accuracy
Fluency
Coherence
Consistency
Extractiveness
Capturing Perspective
Faithfulness
Updated 1mo ago
Evaluation Results
Method
Method
Links
Perspective Accuracy
Fluency
Coherence
Consistency
Extractiveness
Capturing Perspective
Faithfulness
GPT-4
setting=zero-shot
2024.06
93.56
3.63
3.88
3.55
3.38
3.95
3.66
Reference
2024.06
92.65
4.42
4.29
4.21
4.1
4.53
4.75
PLASMA
2024.06
87.27
3.83
3.76
3.62
3.55
3.89
3.98
Flan-T5
2024.06
71.25
3.39
3.7
3.4
3.48
3.76
3.81
Feedback
Search any
task
Search any
task
