FiD+Distil

Reported on 5 benchmarks across 1 task · 1 paper · 1 SOTA

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

Natural Language Processing5 results

Question AnsweringonTriviaQA
EM· uses extra data· 2020-12-08
72.1
best: 87.5 (Claude 2 (few-shot, k=5))
SOTA
Distilling Knowledge from Reader to Retriever for Question Answering arXiv:2012.04584
Question AnsweringonNarrativeQA
BLEU-1· 2020-12-08
35.3
best: 54.11 (Masque (NarrativeQA + MS MARCO))
Distilling Knowledge from Reader to Retriever for Question Answering arXiv:2012.04584
Question AnsweringonNarrativeQA
BLEU-4· 2020-12-08
7.5
best: 30.43 (Masque (NarrativeQA + MS MARCO))
Distilling Knowledge from Reader to Retriever for Question Answering arXiv:2012.04584
Question AnsweringonNarrativeQA
METEOR· 2020-12-08
11.1
best: 26.13 (Masque (NarrativeQA + MS MARCO))
Distilling Knowledge from Reader to Retriever for Question Answering arXiv:2012.04584
Question AnsweringonNarrativeQA
Rouge-L· 2020-12-08
32
best: 59.87 (Masque (NarrativeQA + MS MARCO))
Distilling Knowledge from Reader to Retriever for Question Answering arXiv:2012.04584