Alpaca
Benchmarks
TaskDatasetSOTA resultTrendResults
Alpaca
5.27Speedup (x)
173
Alpaca-P
86.66LLM-as-a-Judge Score
91
Alpaca
3.22Perplexity
61
Alpaca
5.56Speedup
57
Alpaca OnlyTarget Long (malicious)
8.378ATGR
56
Alpaca OnlyTarget Long (benign)
6.115ATGR
56
Alpaca OnlyTarget Short
100TPR
56
Alpaca OnlyTarget Medium
100TPR
56
Alpaca
6.27Average Accepted Length
51
Alpaca
68.8Generation Time (s)
44
Alpaca
0HRR
38
Alpaca SFT Small Prompts
46.6Self-BLEU
36
Alpaca SFT Short Stories
3.7Self-BLEU (Diversity)
36
Alpaca
52.2BRT Score
36
Alpaca AddTarget Medium
100TPR
35
Alpaca clean (test)
79.81F1 Score
32
Alpaca
2.427Perplexity (PPL)
30
Alpaca poisoned (test)
99.93F1 Score
28
Alpaca
0.3Average PTLA (s/token)
28
Alpaca
100TPR
28
Alpaca
40.6BRT Accuracy
26
Alpaca random subset of 1,000 samples
3.19Perplexity
24
Alpaca
32.26Accuracy
24
Virtual-Alpaca
44Forget Rate
24
Alpaca Finance
2.78Average Length
22