Hybrid H3 125M (3-shot, logit scoring)

Reported on 4 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 Processing4 results

Question AnsweringonMultiRC
EM· 2022-12-28
48.9
best: 69.2 (PaLM 540B (finetuned) )
Hungry Hungry Hippos: Towards Language Modeling with State Space Models arXiv:2212.14052
Question AnsweringonBoolQ
Accuracy· 2022-12-28
56.1
best: 99.87 (Mistral-Nemo 12B (HPT))
Hungry Hungry Hippos: Towards Language Modeling with State Space Models arXiv:2212.14052
Word Sense DisambiguationonWords in Context
Accuracy· 2022-12-28
49.1
best: 85.3 (COSINE + Transductive Learning)
Hungry Hungry Hippos: Towards Language Modeling with State Space Models arXiv:2212.14052
Coreference ResolutiononWinograd Schema Challenge
Accuracy· 2022-12-28
43.3
best: 100 (PaLM 540B (fine-tuned))
Hungry Hungry Hippos: Towards Language Modeling with State Space Models arXiv:2212.14052