Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Language Generation on Self-Instruct (test)
Loading...
23.4
ROUGE-L
Teacher
9.464
13.082
16.7
20.318
Mar 4, 2026
ROUGE-L
Updated 1mo ago
Evaluation Results
Method
Method
Links
ROUGE-L
Teacher
Model=LLaMA, #Params=13B
2026.03
23.4
SFT w/o KD
Model=LLaMA, #Params=7B
2026.03
20.8
SeqKD
Model=LLaMA, #Params=7B
2026.03
20.8
VQAE
Model=LLaMA, #Params=7B
2026.03
20.5
KD
Model=LLaMA, #Params=7B
2026.03
20.2
Teacher
Model=GPT-2, #Params=1.5B
2026.03
14.3
SeqKD
Model=GPT-2, #Params=760M
2026.03
14
KD
Model=GPT-2, #Params=760M
2026.03
13.4
VQAE
Model=GPT-2, #Params=760M
2026.03
13.3
SFT w/o KD
Model=GPT-2, #Params=760M
2026.03
12.4
KD
Model=GPT-2, #Params=120M
2026.03
10.8
VQAE
Model=GPT-2, #Params=120M
2026.03
10.4
SeqKD
Model=GPT-2, #Params=120M
2026.03
10.1
SFT w/o KD
Model=GPT-2, #Params=120M
2026.03
10
Feedback
Search any
task
Search any
task