Vega v2 6B (fine-tuned)

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 Processing6 results

Question AnsweringonMultiRC
EM· 2022-12-04
62.4
best: 69.2 (PaLM 540B (finetuned) )
Toward Efficient Language Model Pretraining and Downstream Adaptation via Self-Evolution: A Case Study on SuperGLUE arXiv:2212.01853
Question AnsweringonMultiRC
F1· 2022-12-04
88.2
best: 90.1 (PaLM 540B (finetuned) )
Toward Efficient Language Model Pretraining and Downstream Adaptation via Self-Evolution: A Case Study on SuperGLUE arXiv:2212.01853
Question AnsweringonBoolQ
Accuracy· 2022-12-04
90.5
best: 99.87 (Mistral-Nemo 12B (HPT))
Toward Efficient Language Model Pretraining and Downstream Adaptation via Self-Evolution: A Case Study on SuperGLUE arXiv:2212.01853
Common Sense ReasoningonReCoRD
EM· 2022-12-04
93.9
best: 95.9 (Turing NLR v5 XXL 5.4B (fine-tuned))
Toward Efficient Language Model Pretraining and Downstream Adaptation via Self-Evolution: A Case Study on SuperGLUE arXiv:2212.01853
Common Sense ReasoningonReCoRD
F1· 2022-12-04
94.4
best: 96.4 (Turing NLR v5 XXL 5.4B (fine-tuned))
Toward Efficient Language Model Pretraining and Downstream Adaptation via Self-Evolution: A Case Study on SuperGLUE arXiv:2212.01853
Word Sense DisambiguationonWords in Context
Accuracy· 2022-12-04
77.4
best: 85.3 (COSINE + Transductive Learning)
Toward Efficient Language Model Pretraining and Downstream Adaptation via Self-Evolution: A Case Study on SuperGLUE arXiv:2212.01853