PaLM-540B (Few-Shot)

Reported on 4 benchmarks across 3 tasks · 1 paper · 2 SOTA

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

Natural Language Processing3 results

Question AnsweringonTriviaQA
EM· uses extra data· 2022-04-05
81.4
best: 87.5 (Claude 2 (few-shot, k=5))
SOTA
PaLM: Scaling Language Modeling with Pathways arXiv:2204.02311
Question AnsweringonWebQuestions
EM· 2022-04-05
43.5
best: 84.6 (PoG-GPT4 (Tan et al., 2024))
PaLM: Scaling Language Modeling with Pathways arXiv:2204.02311
Sentence CompletiononHellaSwag
Accuracy· 2022-04-05
83.8
best: 96.1 (CompassMTL 567M with Tailor)
PaLM: Scaling Language Modeling with Pathways arXiv:2204.02311

Medical1 result

Language ModellingonLAMBADA
Accuracy· 2022-04-05
89.7
SOTA
PaLM: Scaling Language Modeling with Pathways arXiv:2204.02311