|
gptkbp:instanceOf
|
gptkb:large_language_model_family
|
|
gptkbp:architecture
|
gptkb:transformation
|
|
gptkbp:author
|
gptkb:Meta_AI_Research
|
|
gptkbp:chatModelVariant
|
gptkb:LLaMA_2-Chat
|
|
gptkbp:context
|
4096 tokens
|
|
gptkbp:developedBy
|
gptkb:Meta_AI
|
|
gptkbp:fineTuningAvailable
|
yes
|
|
gptkbp:hasModel
|
decoder-only transformer
|
|
gptkbp:intendedUse
|
gptkb:research
commercial applications
|
|
gptkbp:license
|
custom commercial license
|
|
gptkbp:notableCompetitor
|
gptkb:PaLM_2
gptkb:Yi
gptkb:Falcon_LLM
gptkb:Gemini
gptkb:Phi-2
gptkb:ChatGPT
gptkb:GPT-4
gptkb:Bard
gptkb:Claude
gptkb:Mistral
gptkb:Vicuna
gptkb:ERNIE
gptkb:BLOOM
gptkb:GPT-3.5
gptkb:MPT
gptkb:Mixtral
gptkb:Qwen
Alpaca
|
|
gptkbp:notableDeployment
|
gptkb:Hugging_Face
gptkb:AWS
gptkb:Microsoft_Azure
|
|
gptkbp:notableFeature
|
improved safety over LLaMA 1
optimized for dialogue
supports RLHF
supports supervised fine-tuning
|
|
gptkbp:notablePublication
|
gptkb:Llama_2:_Open_Foundation_and_Fine-Tuned_Chat_Models
https://arxiv.org/abs/2307.09288
|
|
gptkbp:openSource
|
partially
|
|
gptkbp:predecessor
|
LLaMA models
|
|
gptkbp:releaseYear
|
2023
|
|
gptkbp:size
|
13B
7B
70B
|
|
gptkbp:supports
|
English
code generation
reasoning tasks
instruction following
|
|
gptkbp:trainer
|
publicly available data sources
|
|
gptkbp:trainingTokens
|
2 trillion
|
|
gptkbp:bfsParent
|
gptkb:LLaMA_2:_Open_Foundation_and_Fine-Tuned_Chat_Models
|
|
gptkbp:bfsLayer
|
8
|
|
https://www.w3.org/2000/01/rdf-schema#label
|
LLaMA 2 models
|