RuGPT-3 Medium

Reported on 6 benchmarks across 3 tasks · 1 paper

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Natural Language Processing3 results

Question AnsweringonRuOpenBookQA
Accuracy· 2022-10-23
57.2
best: 86.5 (Human benchmark)
TAPE: Assessing Few-shot Russian Language Understanding arXiv:2210.12813
Question AnsweringonCheGeKa
Accuracy
0
best: 64.5 (Human benchmark)
Question AnsweringonMultiQ
Accuracy
0
best: 91 (Human benchmark)

Methodology2 results

Logical ReasoningonWinograd Automatic
Accuracy· 2022-10-23
57.2
best: 87 (Human benchmark)
TAPE: Assessing Few-shot Russian Language Understanding arXiv:2210.12813
Logical ReasoningonRuWorldTree
Accuracy · 2022-10-23
38
best: 83.7 (Human benchmark)
TAPE: Assessing Few-shot Russian Language Understanding arXiv:2210.12813

Miscellaneous1 result

EthicsonEthics (per ethics)
Accuracy· 2022-10-23
44.1
best: 67.6 (Human benchmark)
TAPE: Assessing Few-shot Russian Language Understanding arXiv:2210.12813