|
gptkbp:instanceOf
|
gptkb:large_language_model
|
|
gptkbp:architecture
|
gptkb:transformation
|
|
gptkbp:availableOn
|
gptkb:Hugging_Face
gptkb:AWS
gptkb:Google_Cloud
gptkb:Azure
|
|
gptkbp:context
|
8,192 tokens
|
|
gptkbp:developedBy
|
gptkb:Meta_AI
|
|
gptkbp:format
|
gptkb:text
|
|
gptkbp:hasModel
|
decoder-only transformer
|
|
gptkbp:intendedUse
|
gptkb:research
commercial applications
|
|
gptkbp:language
|
English
|
|
gptkbp:license
|
custom commercial license
|
|
gptkbp:notableFeature
|
better instruction following
improved reasoning
higher accuracy than Llama 2
|
|
gptkbp:notableUser
|
gptkb:Google
gptkb:Microsoft
gptkb:Hugging_Face
gptkb:AWS
gptkb:Meta
|
|
gptkbp:openSource
|
false
|
|
gptkbp:parameter
|
70 billion
|
|
gptkbp:partOf
|
Llama 3 family
|
|
gptkbp:platform
|
high-end GPUs
|
|
gptkbp:predecessor
|
gptkb:Llama_2_70B
|
|
gptkbp:releaseDate
|
2024-04-18
|
|
gptkbp:supports
|
gptkb:translator
gptkb:dictionary
chat
code generation
question answering
summarization
text generation
reasoning tasks
extraction
multi-turn dialogue
|
|
gptkbp:tokenizer
|
Byte-Pair Encoding (BPE)
|
|
gptkbp:trainer
|
reinforcement learning from human feedback
supervised fine-tuning
|
|
gptkbp:trainingDataCutoff
|
2023-03
|
|
gptkbp:trainingDataSource
|
publicly available data
Meta-licensed data
synthetic data
|
|
gptkbp:bfsParent
|
gptkb:LLaMA_model_family
gptkb:Llama_API
|
|
gptkbp:bfsLayer
|
8
|
|
https://www.w3.org/2000/01/rdf-schema#label
|
Llama 3 70B
|