Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

LongLaMP

Benchmarks

Task NameDataset NameSOTA ResultTrend
Abstract generationLongLaMP
R142.5
32
Personalized WritingLongLaMP Wri
R1 Score30.79
16
Review GenerationLongLaMP Rev
R1 Score34.81
16
Abstractive SummarizationLongLaMP Abs
ROUGE-143.91
16
Topic generationLongLaMP
R130.8
16
Personalized Topic WritingLongLaMP
ROUGE-L23.07
12
Personalized Review WritingLongLaMP
ROUGE-L27.84
12
Personalized Abstract GenerationLongLaMP
ROUGE-L38.92
12
Personalized Long-Form GenerationLONGLAMP Topic Writing (user-based)
ROUGE-10.3139
9
Personalized Long-Form GenerationLONGLAMP Abstract Generation (user-based split)
ROUGE-137.16
9
Personalized Long-Form GenerationLONGLAMP Product Review (user-based split)
ROUGE-136.63
9
Personalized GenerationLongLaMP Pair A Writing (test)
ROUGE-130.79
8
Personalized GenerationLongLaMP (Pair A) - Review (test)
ROUGE-133.03
8
Personalized GenerationLongLaMP (Pair A) - Abstract (test)
ROUGE-141.35
8
Personalized Text GenerationLongLaMP
Alignment Score74
7
Personalized Text GenerationLongLaMP PTW Qwen3-4B (test)
ROUGE-L21.02
6
Personalized Text GenerationLongLaMP PRW Qwen3-4B (test)
ROUGE-L26.68
6
Personalized Text GenerationLongLaMP PAG Qwen3-4B (test)
ROUGE-L37.27
6
Product Title WritingLongLaMP PTW (test)
ROUGE-L22.18
6
Post Review WritingLongLaMP PRW (test)
ROUGE-L27.71
6
Profile Attribute GenerationLongLaMP PAG (test)
ROUGE-L37.56
6
Personalized Text GenerationLongLaMP Product
ROUGE20.3
5
Personalized Text GenerationLongLaMP Topic
ROUGE22.6
5
Personalized Text GenerationLongLaMP Abstract
ROUGE27.1
5
Personalized Text GenerationLongLaMP Average
Personalization Score29.67
4
Showing 25 of 28 rows