List your company

ResearchFormerly WizWand

Generative Task

Benchmarks

DatasetSOTA methodMetricTrendResultsLast updated

PubMedQA labeled (0.8:0.2)

76.8Accuracy

28

Jun 23, 2026

Alpaca random subset of 1,000 samples

3.19Perplexity

24

Jun 5, 2026

28.85Perplexity

6

Feb 26, 2026

Big Patent (test)

21.91Perplexity

6

Feb 26, 2026

28.4BS (Bits Per Character)

4

Feb 26, 2026

T5(3B) + PE w/ ROE (ORC.)

60.02BS Score

4

Feb 26, 2026

T5(3B) + PE w/ ROE (ORC.)

58.36BS

4

Feb 26, 2026

T5(3B) + PE w/ ROE (ORC.)

63.2BS

4

Feb 26, 2026

T5(3B) + PE w/ ROE (ORC.)

52.49BS

4

Feb 26, 2026

T5(3B) + PE w/ ROE (ORC.)

30.16ROUGE Score

4

Feb 26, 2026

T5(3B) + PE w/ ROE

35.55ROUGE Score

4

Feb 26, 2026

T5(3B) + PE w/ ROE (ORC.)

0.316BLEU Score

4

Feb 26, 2026