ELI5
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
ELI5
16.85Answer in Context
7
ELI5 64-bit bitstrings LLAMA3.1-8B
0Message Accuracy
6
ELI5 32-bit bitstrings LLAMA3.1-8B
0.3Message Accuracy
6
ELI5-Category Class Unlearning
42.24Perplexity
6
ELI5-Category Class Unlearning (Drt)
43.27Perplexity
6
ELI5-Category Class Unlearning (Df)
33.51Perplexity
6
ELI5 prompts Gemma-7B-it 200 tokens (test)
1.778Perplexity
6
ELI5
27.9B-1 Score
6
ELI5 (test)
21.3Rouge-L
5
ELI5 (test)
0.038Error Rate
5
ELI5 standard original
26.9RL Score
5
ELI5
5.14Relevance (Mean)
5
ELI5 T5-Large Attack
0.998ROC-AUC
4
ELI5 Clean Text
1ROC-AUC
4
ELI5 ALCE (test)
10.5Correctness
4
eli5
63.2BS
4
ELI5 ALCE (dev)
87.42FSupp
3
ELI5 (test)
29.15ROUGE-1
2
ELI5 (test)
23.37ROUGE-1
2
ELI5 unseen (test)
13.97ROUGE-1
2